Online Safety (Relief and Accountability) Bill

Members may also access these materials through the SG Parl MP mobile app.

Members may refer to Handout 1. We can see that many victims add up to some very disturbing numbers. A Ministry of Digital Development and Information (MDDI) study revealed that more than four in five had encountered harmful online content in the past year. This included content supporting illegal activities, sexual content and violent content. Many of these harms occur on platforms where people go for entertainment, such as Facebook, YouTube and Instagram.

In addition to POHA, the Government has made several moves to enhance online safety. In 2022, we amended the Broadcasting Act to require social media services to disable access to egregious content, including content advocating suicide or self-harm. In 2023, we introduced the Code of Practice for Online Safety. Designated social media services were required to put in place systems and processes to enhance online safety, especially for children and youths. These laws have helped in some measure to stem the tide of online harms. However, they were not a panacea, nor could they be. The impact of technology evolves and is often understood only with time. We have therefore made it a point to review our laws regularly.

Most recently, in January 2024, Parliament debated a Motion on Building an Inclusive and Safe Digital Society. The Motion and other Parliamentary Questions filed by MPs have contributed many useful ideas. They include calls by Mr Darryl David for victim-centric remediation; Ms Mariam Jaafar and Dr Wan Rizal for better platform accountability; Ms Tin Pei Ling, for more timely responses and respectful behaviours online; Ms Nadia Samdin, to lift the veil of online anonymity; Ms Hany Soh to address the use of deepfakes for extortion; Mr Yip Hon Weng, to firmly combat cyberbullying; and Mr Abdul Muhaimin, to deter hate speech.

Mr Speaker, these calls by Members have reinforced where we can do more to strengthen protection for victims of online harms. They align closely with multiple studies conducted over the years by the Government, academics and non-government organisations.

Let me share some important findings from these studies. The first important finding is that victims often want quick takedowns of online harms. A study by the Infocomm Media Development Authority (IMDA) showed that platforms take about five days or more to act on valid reports of online harms. Many valid reports are not even acted on. This is highly unsatisfactory to victims.

The second important finding is that victims are daunted by the existing channels for seeking remedies and in particular, they find court processes complex and expensive. The third important finding is that victims face difficulties seeking restitution. This is because they are often unable to confirm who was responsible for the harm. They continue to be in fear and cannot find closure.

Today, we are taking a decisive step to introduce legislation. It will give victims of online harms an avenue of timely, effective and accessible redress. OSRA aims to do this in three ways.

First, OSRA will establish the Office of the Commissioner of Online Safety. An agency, the Online Safety Commission (OSC), will be set up to support the Office of the Commissioner. OSC will be empowered to issue directions to communicators of harmful content, administrators of groups or pages and platforms. These directions require the recipients to act quickly on the harmful content. OSC will start by tackling online harms that are most prevalent and Singaporeans consider the most severe. Other harms will be dealt with progressively. The Minister of State Rahayu will elaborate on how OSC is intended to operate.

Second, OSRA will introduce statutory torts. This is a major improvement. It provides a legal basis for the victim to take action. Most importantly, it will allow victims to initiate legal proceedings against those responsible for causing or continuing the harm.

Sir, studies have shown that Singaporeans see online safety as a shared responsibility between technology companies, individuals, the Government, parents and schools. The proposed statutory torts will clarify the duties that users, administrators of groups and platforms owe to each other. It will significantly reduce ambiguity about what online behaviours are considered harmful and unlawful.

Third, OSRA provides victims with the means to identify the person who caused the harm. Not all victims want this, but for those who do, OSRA plugs a gap. Minister Edwin Tong will elaborate on this.

Sir, if Parliament passes OSRA, Singapore will be one of only a few countries worldwide to have an agency dedicated to helping victims of online harm. While we have few counterparts to learn from, this does not deter us. In setting up the OSC, we considered various models, such as mediation and dispute resolution.

Victims are at the very heart of the OSC's mission. But the OSC's operational model and its oversight process are designed to be reasonable and fair to all. This is balanced against the need to act expeditiously and reduce the possibility of victim re-traumatisation. We also need to minimise long-drawn investigations and appeal cases.

In designing our set-up for the Office of the Commissioner of Online Safety, we drew lessons from Australia's eSafety Commissioner. We learnt much from their generous sharing of experience to build a model that works for Singapore. For example, eSafety started small, focusing only on image-based child abuse material and cyberbullying of a child or young person under 18.

To put OSC on a firm footing, we will, similarly, also start off by tackling the most severe and prevalent harms. Minister of State Rahayu will share more on the phased implementation approach.

While we learnt important lessons from Australia's eSafety, I should point out, there are differences. For example, the Commissioner of Online Safety in Singapore will be empowered to deal with a wider set of harms. OSRA identifies thirteen types of harms which affect Singaporeans from all walks of life.

The Commissioner can issue directions other than what is colloquially known as a "takedown". For instance, it can issue a Right-of-Reply Direction that will allow the offending content to be seen alongside the victim's reply. The Commissioner can also seek the platform's cooperation to obtain identity information of perpetrators.

Sir, I cannot overstate the importance of the OSC being able to act swiftly and effectively. This victim-centric approach is the cornerstone of the OSC. After all, what Singaporeans have said they want most is a simple and effective way to protect themselves from online harms.

Importantly, it sends a signal that harms to victims inevitably become harms to society. Our collective well-being is compromised when those who are harmed are denied restitution. With online harms becoming more prevalent, our barometer for acceptable online behaviour has been steadily eroded. But just as we would not tolerate harmful behaviours in the physical world, we also cannot allow bad behaviours to become normalised online.

Through OSRA, we hope to avoid such an eventuality. In its place, we want to lay the foundations for our citizens' online interactions by fostering trust in online spaces, Singaporeans can participate safely and confidently in our digital society.

Sir, we will continue to work with all our stakeholders from technology companies to community partners to implement OSRA. We continue to welcome Members' views on how we can further improve online safety for Singaporeans.

Mr Speaker, may I have your permission to provide a summary in Mandarin, please?

(In Mandarin): [Please refer to Vernacular Speech.] Harmful content online has seriously affected our online environment and goes against our social values.

The Online Safety (Relief and Accountability) Bill focuses on three main areas to strengthen online safety and protect Singaporeans from online harm.

First, establishing the OSC will enable victims to receive help more quickly. When harmful content is encountered, the OSC can also issue directions requiring perpetrators or platforms to swiftly remove relevant content and stop its dissemination. This is precisely the outcome that most victims wish for. Second, the Bill also clearly lists out unlawful online behaviour, allowing victims to take civil action against the perpetrators and make perpetrators bear legal consequences for their actions. Finally, the OSC can help victims obtain the identity information of perpetrators. Although perpetrators often hide their identities, we are discussing with various platforms how to ensure that perpetrators cannot hide forever.

Overall, the purpose of the Bill is not only to provide timely assistance to victims, but also to help them seek justice. We also hope the Bill can serve as a deterrent, set new norms for online behaviour and create a safer, more positive online space.

(In English): Sir, our journey towards building a safer online world does not stop with OSRA. We are especially concerned about younger persons and plan to do more to ensure the usage of online services is age-appropriate. We know anecdotally and from studies that youths can easily bypass age restrictions online. They use false birthdays, spam email accounts or borrowed credentials for verification. When one person figures out how to do so, they teach their friends, too.

Such behaviours can cause more youths to be exposed to inappropriate content or persons. To moderate such risks, we have introduced age assurance measures on app stores. Hopefully, young users do not download age-inappropriate apps in the first place, but we know this is not enough. A more holistic approach is needed.

This means extending age assurance requirements to social media services which attract many young users. We will explore better protections on other platforms, such as online games, where children and youths spend much of their time.

In addition, we will look beyond content to address unwanted interactions and excessive time spent on apps. We will review the additional measures needed to better protect younger users online. As and when new technology arises, we will study them and update our safeguards, where required.

Mr Speaker, let me conclude. In the debate to follow, I suggest Members keep in mind three points.

One, OSRA introduces many new features in our online safety ecosystem. They are novel and meaningful. Nevertheless, there is much work ahead to ensure that these features can be effectively implemented.

Two, OSRA is designed to help victims in areas and ways that matter most to them. This was how we determined the types of harms, range of directions and thresholds of harms. We have also introduced safeguards to prevent overreach of the Commissioner and OSRA.

Three, these considerations should, therefore, inform further proposals and suggestions. We should be careful not to distract the Commissioner and the OSC, or impose more conditions which will impede victims from getting timely, effective and accessible relief.

With that, Mr Speaker, I look forward to the debate on this Bill. Sir, I seek to move.

*Question proposed.*

Mr Speaker, earlier, Minister Josephine Teo spoke about the genesis, the intent of the OSRA Bill and the need for more to be done to improve online safety. I will take Members through the Bill.

Let me first give an overview of the structure of the Bill. The Bill is split into 15 Parts. Part 1 consists of standard provisions, such as the purpose and interpretation sections. Part 2 establishes the Office of the Commissioner of Online Safety, or the Commissioner. Part 3 defines the categories of harm that will fall under the scope of the Bill. Part 4 governs the statutory reporting mechanism; and Part 5, the types of directions that may be issued by the Commissioner. Part 6 covers the investigative powers of the Commissioner. Part 7 governs the oversight mechanisms that the Commissioner will be subject to. Part 8, on offences and enforcement. Part 9 consists of provisions relating to service of documents and powers to make regulations. Parts 10 to 13 establish the statutory torts and the damages that may be sought under the torts, which Minister Edwin Tong will speak on later. Part 14 deals with jurisdictional, procedural and miscellaneous matters, and Part 15 covers amendments made to other Acts.

In my speech, I will be focusing on Parts 2 to 8 of the Bill, and will explain how the Bill will help to define what behaviours are unacceptable in the online world.

Part 3 of the Bill will specify 13 categories of harm, with each category defined in legislation, to create a common understanding of the scope and definition of each type of harm. These definitions were developed in consultation with the public, platforms and were guided by a Steering Committee. The Steering Committee was chaired by then-Minister for Law and Home Affairs, Mr K Shanmugam, and made up of Government agencies and experienced industry members.

Let me now explain each harm in turn.

First, online harassment, including sexual harassment. Online harassment is defined as the communication of online material that a reasonable person would conclude is threatening, abusive, insulting, sexual or indecent, and is likely to cause a person harassment, alarm, distress or humiliation.

Second, doxxing, which will cover the publication of a person's identity information that a reasonable person would conclude is likely to have been intended to cause harassment, alarm, distress and humiliation.

Third, online stalking, which will be defined as engaging in a course of online conduct that involves online acts or omissions associated with stalking, and that a reasonable person would conclude is likely to cause harassment, alarm, distress or humiliation to the other person.

Fourth, intimate image abuse or the non-consensual sharing of intimate images, a harm that can result in a severe and lasting impact on victims. This is defined as the communication of online material that contains an intimate image or recording of a person, without their consent, that a reasonable person would conclude is likely to cause that person harassment, alarm, distress or humiliation. This includes images or recordings that may have been altered or generated by artificial intelligence (AI) or any other means. An offer to sell or distribute, and advertisements of an intimate image or recording, are also included.

Fifth, image-based child abuse. Like intimate image abuse, this will include images or recordings that may have been altered or generated by AI or other means and will cover depictions of both physical and sexual child abuse.

Sixth, inauthentic material abuse. With the rapid development in AI technology and the advent of more powerful and easily accessible generative AI models, we have seen more deepfakes being used to cause harm to others. This harm is defined as the communication of inauthentic material that a reasonable person would conclude is likely to cause harassment, alarm, distress or humiliation because it is false or misleading.

The material is regarded to be inauthentic where it is an image, video or soundbite that has been manipulated or generated through digital means to create a false or misleading depiction of the victim's words, actions or conduct, and is realistic enough such that a reasonable person would believe that the victim did say such words or engaged in such actions or conduct. This category includes not only expressed depictions but also implied depictions.

Seventh, online impersonation, which is defined as online activity where one person pretends to be another without their consent and would lead a reasonable person to believe that the activity is conducted by the second person. However, parody, satire or commentary which no reasonable person would believe is made by the victim, is not online impersonation.

The eighth category is online instigation of disproportionate harm. This harm seeks to address the issue of "mob behaviour" or cancel campaigns, where one individual calls for numerous others to pile on to another person, in response to the person's actions or speech, instigating disproportionate harm. This category of harm will not be applicable to statements that tend to rally or call the public to undertake core political activities, such as calling for others to vote for a candidate in a Presidential or Parliamentary election.

The ninth category is the non-consensual disclosure of private information. This harm is defined as the publication of any private information of a person without consent, and that a reasonable person would conclude was likely to cause that person harassment, alarm, distress or humiliation. The types of information that would be deemed as private would be context specific, and would depend on whether the information is already in the public domain. Whether the individual is of the view that the information is sensitive may also be taken into account by the Commissioner. The Minister for Digital Development and Information will also be able to prescribe certain types of information as "private information".

The 10th and 11th categories are the incitement of enmity and the incitement of violence. Enmity refers to feelings of enmity, hatred or hostility against any group. Violence refers to unlawful force or unlawful violence. These harms are what we term "group harms", as they aim to tackle harmful online content that is likely to harm groups in Singapore. A group will be defined as a group of persons of any description and may include, for example, a group distinguished by race or religion.

Lastly, the 12th and 13th categories are the publication of false material and the publication of statements harmful to reputation. The former will cover statements about a person that are false and that a reasonable person would conclude are likely to cause harm. The latter will cover statements that a reasonable person would conclude are likely to harm the reputation of the victim and likely to cause any other additional harm to the victim. These two categories of harm differ slightly from the others in terms of the remedies victims may seek. I will touch more on this later.

There have been suggestions to include two new categories, namely sexual grooming and the publication of online material that encourages or promotes suicide or acts of self-injury. These harms have not been included at this stage, because, firstly, these harms are already addressed through other measures in our broader online safety framework, through the Penal Code for sexual grooming, and the Broadcasting Act for online material that encourages or promotes suicide or acts of self-injury.

Secondly, the OSRA Bill is designed in a way that seeks to stop online harm from occurring, such as through the removal of content. This is made possible because the online harm is identified through a single, discrete post, like the 13 categories of harms I described earlier. In contrast, sexual grooming often occurs over the course of communication or exposure to various pieces of content, where each individual piece of content may not be harmful in and of itself.

Thirdly, OSRA is designed in way that relies on user reports. Recourse can only be provided in cases where the victim recognises that they are victims. Unfortunately, it is often the case that victims of sexual grooming or consumers of self-harm and/or suicide content, are not aware that they are victims. However, in such cases where these victims do come to the Commissioner for assistance, the Commissioner will work with the relevant agencies, including the Police, to provide the necessary assistance.

Members might observe that not all of these 13 categories are new and some of them are already being addressed by existing law. For example, online harassment is already being covered by POHA. The Penal Code as well as the Criminal Law (Miscellaneous Amendments) Bill, which was debated in Parliament earlier, both address intimate image abuse. Such content might also be covered as egregious content under the Broadcasting Act. The suggested additions to the categories of harm would also be covered by the Penal Code and the Broadcasting Act.

The group harms have also already been addressed by the Maintenance of Religious Harmony Act and the Maintenance of Racial Harmony Act, for groups distinguished by religion or race. The common law action of defamation, as supplemented by the Defamation Act, provides individuals with an avenue of recourse for defamatory statements or statements which harm a person's reputation. OSRA will complement these existing laws.

Some laws address harms that affect the public interest, such as the Maintenance of Religious Harmony Act, the Maintenance of Racial Harmony Act and the Broadcasting Act. Others provide individual remedies through legal action in Courts, such as POHA. However, victims have shared that seeking legal remedies is often a lengthy and expensive legal process, which often deters them from seeking recourse.

This can be seen from the SHE's 2023 Online Harms Report, which showed that 28% of respondents who decided not to take legal action did so due to the cost. The same SHE report also found that most respondents preferred the swift and permanent removal of content over taking legal action. OSRA is intended to complement these existing legislations by expanding the scope of harms covered, through including new harms, such as inauthentic material abuse and non-consensual disclosure of private information, and by allowing victims to seek recourse in a simple and timely manner.

We acknowledge that the Internet evolves rapidly and the online harms ecosystem may change drastically over short periods, with new harms emerging or with bad actors finding new ways to cause harm. Thus, OSRA has been designed in a way to allow us to adapt to changes in the online harms ecosystem.

The Minister for Digital Development and Information will be able to prescribe additional types of online harms. This power will be vital in ensuring that we address emerging harms as soon as possible. We will ensure that this power is used judiciously and for harms that are particularly egregious to individuals, to prevent the unchecked expansion of the scope of OSRA.

Part 2 of the Bill establishes the Office of the Commissioner of Online Safety. It will be responsible for administering the statutory reporting mechanism to provide timely relief for victims of online harms. The Commissioner of Online Safety will be appointed by the Minister for Digital Development and Information and will be supported by Deputy Commissioners and Assistant Commissioners.

While the Minister may provide broad guidance to the Commissioner, the Commissioner will be the final decision-maker for all cases. We will also establish a new agency, called the OSC, that will support the Office of the Commissioner for Online Safety. The Commissioner may delegate the exercise of all of, or any of, the functions or powers of the Commissioner to the officers of the OSC, to allow them to effectively support the Commissioner. This includes the power to issue directions, which I will speak more on later.

However, the Commissioner may not delegate the power of appointment or delegation, or the power to issue advisory guidelines. The OSC will be administratively supported by the IMDA.

As announced by Prime Minister Lawrence Wong during his speech at the Smart Nation 2.0 Launch in October 2024 and reiterated by Minister Josephine Teo earlier in March this year, the OSC will be set up in the first half of 2026. We have also considered the importance of transparency. To that end, the Commissioner will consider publishing regular reports for public awareness on online harms and the Commissioner's work, which may include information on aggregated caseloads, and anonymised information, insofar as these do not re-traumatise victims.

Let me now say more about the reporting process, which, as I shared earlier, is governed by Part 4 of the Bill. With your permission, Mr Speaker, may I ask the Clerk to distribute a set of handouts to the Members?

Members may also access these materials through the MP@SGPARL mobile app. Members may refer to Handout 2 for an overview of the user journey with the OSC.

First, prior to submitting a report to the OSC, victims will, generally, be required to report the online harm to the platforms. While we are setting up the OSC to provide timely relief to victims of specified online harms, platforms remain the first port of call. Platforms must continue to play an active role and take responsibility for the safety of their users online. Where platforms fail to act on online harms within 24 hours, victims can then file a report to OSC.

The requirement to first report to platforms before reporting to the OSC will, however, not be required for certain egregious harms, such as intimate image abuse and image-based child abuse. This step will also not be required for doxxing. Victims of these harms can submit a report directly to the OSC.

Our view has always been that platforms should take responsibility for keeping their users safe online. This is evident from the approach we have taken under the Broadcasting Act. For example, under the Code of Practice for Online Safety – Social Media Services, designated social media services (DSMSs) with significant reach or impact in Singapore are already required to submit annual reports to be published on IMDA's website. These reports contain information on the DSMSs' measures to combat harmful and inappropriate content, and metrics, such as the number of reports received from users, as well as the response times to act on these reports.

There are other baseline requirements. To be eligible to submit a report, victims must be Singaporean Citizens, Permanent Residents or have a prescribed connection to Singapore. We intend for this to cover foreigners who are residing in Singapore for the long-term at the onset.

Where victims are under the age of 18, the parent or guardian of the victim may also submit a report on their behalf. For example, earlier in November 2024, we saw deepfake nude photos of students at the Singapore Sports School being created and circulated by the student-athletes. In such a case, both the student victims and the parents of the victims depicted in the deepfakes would be able to submit a report to the OSC.

We have also considered situations where a victim strongly prefers that the report be filed for them. In such situations, victims will be able to authorise other persons to file reports on their behalf. This includes authorising employers or public agencies. For example, a hospital, if authorised, may file a report on behalf of a healthcare worker who is a victim of a specified online harm. In the spirit of the reporting mechanism, OSC will only assess reports submitted. OSC will not actively monitor and identify cases of online harms.

Reports will be submitted through the OSC's website, which will be designed with the user in mind. We are consulting other agencies and third-party organisations, such as SHE, in designing the website and the reporting form, to take into account user-centric language.

Accounting for victims' needs early is important, as survivors of online harms may, sometimes, avoid seeking external support to avoid re-traumatisation. The OSC's website will also provide victims and other persons with access to resources on online harms and advice on how to keep themselves safe online. This will include, for example, information on the different types of online harms and what to do when you have experienced them.

The OSC is also exploring partnerships with third-party organisations which victims may be referred to for support, such as counselling services or further resources. The details of this are still being worked out and we will provide more details in due course.

Each report will be assessed on its own merits and the OSC will be able to act on the face of the report submitted. This is to ensure that the OSC can move fast to address the online harm. Where necessary, such as where the information provided in the report is unclear, the Commissioner will be empowered, under Part 6, to conduct investigations, to better identify the relevant facts of each case. The OSC will also develop internal practices and ensure that case officers are trained to handle each case sensitively.

Let me turn now to directions, which are listed in Part 5 of the Bill.

The Commissioner will be empowered to issue directions to stop online harms from continuing to occur or to prevent further online harms from affecting the victim where there is reason to suspect that the online harm was conducted in respect of the victim or the victim group, as the case may be. The threshold for the issuance of directions is modelled after the threshold set for the Police to take protective action in the Criminal Procedure Code and the Online Criminal Harms Act (OCHA).

We had considered raising the threshold for the issuance of directions to "reasonable grounds to believe". However, we ultimately landed on "reason to suspect", to ensure that online harms can be stopped in a timely manner. We will be putting in place oversight mechanisms to ensure that OSC directions are only issued where appropriate. I will speak more on the oversight mechanism later.

The OSC may issue directions to three different parties: the communicator of the online harm, the administrator of a group or location where the online harm occurred, and the platform on which the online harm occurred.

Broadly, communicators and administrators may be issued directions that require them to remove specified material or disable a specified location, to restrain them from posting certain types of content or carrying out certain types of conduct or to require them to put up a victim's reply.

Administrators may also be issued other directions, such as directions that require them to put up a label to warn visitors that the online location has been subject to previous OSC directions or to restrict access by a Singapore account to the online location managed by them.

Platforms may be issued directions that require them to prevent end-users in Singapore from accessing specified content or online locations, to restrict interactions between an account and end-users in Singapore, to ban a Singapore account or to post a victim’s reply on its service.

As I mentioned earlier, two of the categories of harm – publication of false material and publication of statements harmful to reputation – differ slightly in terms of remedies that victims may receive. Generally, victims of these harms will only be able to seek a Right-of-Reply Direction to allow them to have their side of the story heard. We will be introducing new types of directions requiring recipients to act on content that can be distinguished by unique identifiers, such as usernames, keywords or hashtags.

The OSC may also issue what we consider as Enhanced Directions to certain prescribed online services. These Enhanced Directions impose additional requirements on the prescribed online service and may require the recipient to act on identical online harmful material, to take further steps to prevent online harm from occurring in the future or to reduce engagement of its users with a class of material. Members may refer to Handout 3 for the full list of directions that the OSC may issue, and Handout 4 for illustrations of online harmful activity and possible OSC directions.

In deciding whether to issue a direction and the type of direction to be issued for a case, the OSC may consider a basket of factors, including: the degree of the harm caused or likely to be caused; the number of persons harmed or likely to be harmed; the manner and circumstances in which the online harmful activity occurred; whether the conduct of the online harmful activity was reasonable, such as when a comment or post would be considered to be a "fair comment"; the likelihood of further online harmful activity being conducted; and whether the direction would be contrary to any public interest. These factors provide the OSC with much needed flexibility, to ensure that appropriate action is taken in every case.

The OSC will also publish guidelines detailing the factors that the OSC will consider in its decision-making process. Such guidelines will also include illustrative examples of when the OSC will or will not act.

In cases where the OSC is made aware of non-compliance with its directions, the OSC may take further escalatory actions. These actions include the issuance of orders following non-compliance, such as: access blocking orders, which may be issued to providers of Internet access services, to disable Singapore end-users' access to an online location; or app removal orders, which may be issued to providers of app distribution services, to remove the specified app from the Singapore app stores. As these orders may affect all users in Singapore since they will no longer be able to access the apps or websites, they may only be used after careful consideration.

Part 8 of the Bill provides for offences and enforcement. The directions and orders issued by the OSC are legally binding, and non-compliance with these directions and orders will be a criminal offence. Where there has been non-compliance, the Commissioner will be empowered to conduct further investigations, such as requesting information or documents from individuals. These are the same Part 6 investigative powers that the Commissioner may exercise when assessing reports in the first instance.

In developing the penalties under the OSRA Bill, we have referenced other existing legislation, such as the Broadcasting Act, OCHA and the Penal Code. For example, the penalty for non-compliance with directions will be a fine of up to $20,000, or imprisonment for a period not exceeding 12 months, or both, for individuals. Individuals will also be subject to a continuing fine of up to $2,000 for each day the offence continues after conviction. Entities will be subject to a fine of up to $500,000 and a continuing fine of up to $50,000 for each day the offence continues after conviction.

The provision of false information to the Commissioner or the OSC will also be an offence. Individuals found guilty of this offence will be subject to a fine of up to $20,000, or imprisonment for a period not exceeding 12 months, or both. Entities will be subject to a fine of up to $50,000.

In some cases, it may be more appropriate to focus on rehabilitating the perpetrator rather than prosecuting them for non-compliance with an OSC direction. To that end, the Commissioner may put in place an online harms remedial initiative. This initiative could include the completion of volunteer programmes by the perpetrator, which may be taken into account when considering prosecution for non-compliance with the OSC’s directions.

As I mentioned earlier, we will be putting in place oversight mechanisms, which will be governed by Part 7 of the Bill. These mechanisms ensure that the OSC will be able to move quickly and with confidence on the face of the information it has received.

Victims, recipients of directions or orders, and other prescribed persons will have access to a two-step appeal process. Eligible persons may first apply to the Commissioner to reconsider the OSC’s decision. Thereafter, eligible persons may appeal against the OSC's reconsidered decision to an independent appeal panel that will be appointed by the Minister for Digital Development and Information.

For the reconsideration process, the OSC will re-assess the relevant case afresh, and may also take into account any new information that may be presented by the relevant parties after the initial assessment was conducted. Applicants will be able to submit such new information to the OSC. When a decision has been made, the OSC will inform the applicant and other affected parties of their reconsidered decision. The OSC may affirm, revoke, vary or substitute any earlier decision, direction or order issued.

The appeal panel will consist of individuals from academia, society and industry, across different areas of expertise, and will focus on assessing whether a specified online harm had occurred and whether the reconsidered decision made by the OSC is proportionate and justifiable. Where the victim, recipient or prescribed person is dissatisfied with the OSC’s reconsidered decision, they may submit an application to the appeal panel to appeal against the OSC’s reconsidered decision. Similar to the reconsideration process, the appeal panel will be able to affirm, revoke, vary or substitute decisions of the OSC in relation to the issuance or non-issuance of a direction. The appeal panel will also be able to hear appeals on the issuance of an order following non-compliance.

Each individual will be given one chance to have their case heard by the independent appeal panel. Should the individual continue to be dissatisfied with the outcome of the appeal, they may seek judicial review.

The establishment of a new office and the statutory reporting mechanism is a monumental task, one that will require close collaboration across Government. We have to do the OSC right, so that we can do right by the victims. This is why we will be implementing the reporting mechanism in phases. What this means is that we will be bringing the first five harms I shared about earlier into force within the first six months after the OSC opens its doors. These are: intimate image abuse; image-based child abuse; online harassment, including sexual harassment, doxxing and online stalking. The rest of the harms will follow progressively.

These five harms are prioritised for a reason. They represent the most prevalent and serious harms faced by Singapore users online. These are also the online harms that Singaporeans are the most concerned about. For example, The Institute of Policy Studies’ 2025 Online Safety Study showed that targeted harassment was seen by Singaporeans as “highly harmful”. Members may refer to Handouts 1 and 5, for a broader overview of recent survey results and studies relating to online harms. This will allow us to better manage the OSC's caseload and to properly develop the necessary guidelines, frameworks and capabilities to ensure that the OSC's decisions are consistent and appropriate in all cases. Mr Speaker, allow me to say a few words in Malay.

(In Malay): [Please refer to Vernacular Speech.] The need to provide accessible and timely programmes and resources to victims of online harms is not new. Over the years, we have taken steps to educate and better empower our citizens to act against online dangers. For example, in 2021, we established the Sunlight Alliance for Action (AfA) to tackle online harms.

In the span of one year, AfA organised campaigns to raise public awareness about online harms and their impact. These campaigns also equipped youths to provide better support to their peers who were affected by online harms.

As co-chair of this AfA, I was informed by our partners about the experiences that many victims went through. This is what drives me, to this day, to build a safer digital space.

The Sunlight AfA also inspired many of its members to continue their noble efforts to help victims. One example is the establishment of SHECARES@SCWO, a collaboration between SHE and the Singapore Council of Women's Organisations (SCWO). It is Singapore's first one-stop support centre for victims of online harms.

Efforts to enhance online safety require the involvement of every level of society, and the Government is ready to step forward to strengthen these efforts. With the establishment of the OSC, we will be able to provide timely follow-up actions for victims, help stop harms as quickly as possible and build a more robust support ecosystem for victims.

(In English): Mr Speaker, the OSRA and the OSC are just the one of the many steps that we are taking to enhance online safety. I am heartened to note that online safety is a matter that both sides of this House are passionate about and I invite all Members of this House and the public to continue our conversations on how we can better enhance online safety.

Mr Speaker, Sir, today, we take an important step to make our online spaces safer and fairer. The Bill before this House clarifies duties, provides relief and also strengthens accountability across the online ecosystem to deal with online harms. My colleagues have earlier taken Members through the key aspects of the Bill, and I will focus on outlining the new Statutory Torts Framework, as well as the End-User Identification measures in this Bill.

Sir, let me start by reiterating why we saw a need for this Bill. The victims of online harms are not just statistics. They may be our children, classmates, colleagues or neighbours. Their confidence, studies and indeed their livelihoods can be shaken and seriously affected.

For one Primary 4 student, cyber-bullying reared its ugly head when she started using Instagram. She based her self-esteem on how many "likes" and "followers" she had. She would ask her friends to "like" her photos or "follow" her account to appear popular, and if a photo did not receive over 100 "likes", she would delete it.

She began receiving comments about her appearance and hurtful messages from those who were supposedly her friends. Her mental health worsened. She self-harmed and she was later diagnosed with Post-Traumatic Stress Disorder (PTSD) when she eventually sought help at the Institute of Mental Health (IMH). She was on medical leave from school for most of the year.

Sir, this is a sad story. But what is worse is that it is not an isolated case. With your permission, Mr Speaker, may I ask the Clerks to distribute a handout?

Sir, Members may also access these materials through the MP@SGPARL app.

Sir, Handout 6 sets out several other accounts: students, women, working adults and their loved ones. Online harms affect not only individuals and their families, but also the confidence and the trust of our society as a whole.

In the SHE 2023 Study on Online Harms cited by Minister of State Rahayu a short while ago, 76% of respondents were not comfortable expressing their personal views on potentially controversial issues and topics online. Women, youths and minorities are especially vulnerable: 22% of female youths experienced sexual harassment compared to 14% across all respondents; 52% of respondents aged 15 to 24 years old personally experienced online harms compared to 38% across all respondents; 14% to 21% of respondents believed they experienced harms because of their identity, such as race and religion.

As a result, victims exit from public life or they choose to stay offline. Their voices, and their representation, are lost. And proper discourse in the community becomes weaker. If this continues, we will have a divided society, and a weakening of our social fabric and collective trust.

Sir, in contrast, most Singaporeans feel safe on our streets. About 98% of Singaporeans feel safe walking out alone, even at night. In the offline world, we are mindful of how we behave towards one another. We know not to threaten, harass or insult other people. And in our everyday interactions, there are social norms that we know to abide by. In public spaces, schools and also in work spaces. Such norms allow us to function and thrive as a society.

But when it comes to the online world, these norms are not quite observed in the same way. The sense of safety we have offline does not translate into the online space. The SHE 2023 study found that 58% of respondents reported personally experiencing and/or knowing others who faced online harms.

We studied why this might be so. And in the course of our study, three issues stood out: first, fragmented standards and weak accountability in the online harms regulation space; second, economic incentives that unfortunately reward sensational content; and third, online anonymity that emboldens misbehaviour.

These three issues have very much shaped the thinking behind the construct and the framework of this current Bill. Let me elaborate that for Members.

Thus far, the development of the rules and norms of the Internet has been largely left to the tech companies. Left on their own, different platforms will apply different rules, largely shaped by their own interests.

And in the absence of any common set of enforceable norms, it is not easy to expect or enforce accountability. Instead, these differing standards will continue to exist and operate in a manner which allows wrongdoers to exploit the gaps. This lack of accountability is exacerbated by the structural features of the Internet. There is a misalignment between tech companies' profit motive and the need and desire to enhance online safety.

At the Online Harms Symposium co-organised by the Ministry of Law (MinLaw) and the Singapore Management University (SMU) Yong Pung How School of Law in 2023, former Facebook employee turned whistle-blower, Ms Frances Haugen spoke about how tech companies do have solutions to address online harms, but implementing them will eat into their profits and hurt their bottom line. Social media platforms profit from the amplification of sensational and inflammatory content. Viral, harmful content draws engagement, higher engagements draw eyeballs so their removal will impact revenue.

We have some basis to believe that online safety considerations might well yield to profit generation, if left unregulated. Today, tech companies have community standards which ostensibly address online harms. Members can see some examples of these standards which are set out in Handout 7. However, most platforms do not necessarily adhere to their own community standards. Indeed, in the IMDA study cited by Minister Josephine Teo a short while ago, over 50% of legitimate user complaints were not addressed in the first instance. And this, we believe, is far from ideal, as victims are reliant on platforms to stop online harms.

At present, there is no framework which can give redress to victims for harms that happen at the speed and the extent to which they happen online. While victims may try and seek relief in court, there is a limit to how fast court proceedings can be and may also be costly. And we all know that in these types of cases, speed of redress is crucial. And most victims of such harms do not want to have to seek relief in Court, through the Court process.

It is also clear that online users behave on the Internet differently from the offline world. Perpetrators are emboldened to act with impunity especially when they can remain anonymous.

First, they experience what we call the "Online Disinhibition Effect", a term coined by Prof John Suler. Prof Suler explains that anonymous Internet users separate their online persona from their in-person identity and they do not feel responsible for their behaviour online.

Second, there is little that victims can do to hold anonymous users accountable, simply because they remain anonymous and perpetrators know that they are unlikely to get caught and do not fear consequences.

Overall, in general, anonymity fuels bad behaviour.

To address these issues, the Government has taken proactive steps. Minister Josephine Teo outlined the reforms that protect users. Minister of State Rahayu explained how the new OSC will provide timely relief to stop harm. Let me address the remaining mechanisms in this Bill for Members.

Sir, we start with the proposition that, in some cases, stopping the harm alone might not be enough. Victims might require additional recourse, for example, compensation. And to do that, they will also need to know who lies behind that anonymous social media handle.

The Bill therefore seeks to introduce a framework to close those gaps. It will introduce clear statutory duties for online actors, defining what responsible conduct and therefore, what the standard of duty that is expected of them, ought to be. It provides civil remedies for breaches of those duties, giving victims the right to seek justice in Court and it creates an additional avenue of relief, complementing the quick administrative recourse that can be obtained through the OSC, so that victims can choose the path that best suits their needs.

During the Online Harms Symposium that I referred to a short while ago, Ms Haugen likened Internet and social media regulation to the evolution of road safety, something that happened several decades ago. In the 1960s, US car manufacturers vigorously resisted safety reforms. But informed legislators persisted with greater regulation, coupled with a push, a very determined push from concerned citizens as well as investors.

Through legislative changes, safety was made a core design principle in the manufacturing of cars. And it has been estimated that between 1960 and 2012, over 600,000 lives were saved in the United States (US) as a result of this.

In a similar way, not exactly the same, but in a similar way, we hope that this Bill will make online safety a design principle of the online space upfront, and not just an afterthought.

Sir, let me now outline how the Statutory Torts Framework is intended to work.

At the outset, our focus and intent is to empower victims. It is a very victim-centric approach and today you look at the example of one victim who tries to speak to the platforms, you heard the statistics that Minister Josephine Teo cited earlier, it is very difficult, and they are pretty much powerless today.

At the Government level, we have the Broadcasting Act, we have the Protection from Online Falsehoods and Manipulation Act (POFMA), we have OCHA that empower the Government to act in a variety of ways. OSRA provisions will empower private persons to obtain relief, and Members might, therefore, be aware that clause 4 makes it clear that a public agency cannot commence a claim under the Statutory Torts Framework. This is a private citizens' private remedy.

I will now explain the types of online harms covered by the statutory torts. This framework will cover the same categories of harms that the OSC will act on, the same categories that Minister of State Rahayu took you through earlier, with some exceptions and refinements for coherence.

Clauses 83 to 88 cover the following harmful online activities: intimate image abuse; image-based child abuse; online impersonation; inauthentic material abuse; online instigation of disproportionate harm; and incitement of violence.

Harassment, doxxing and stalking will continue to be dealt with under POHA for communicators, but the new statutory duties for these harms will extend to administrators and platforms under this Bill, since POHA does not cover them.

The online harms omitted from the statutory torts, mainly, false material, statements harmful to reputation and non-consensual disclosure of private information – are already well-covered under the existing legal framework, laws on defamation as well as on privacy and confidentiality.

This alignment ensures coherence in our legislation – no overlapping and no double remedy.

In this Bill, the statutory torts will also not cover Incitement of Enmity. We think it is unwise to encourage such matters – which can be potentially explosive, emotive and divisive – to be dealt with litigiously, in a courtroom and so it will be dealt with by the OSC.

The statutory torts will be implemented in phases, as you heard Minister of State Rahayu sketch out the OSC’s implementation of harms framework earlier in coordination with the OSC.

Second, Sir, the Bill assigns clear duties to the key actors in the online ecosystem, and they are the communicators, the administrators and the platforms.

Clauses 83 to 88 therefore impose duties on communicators not to make or share any communication which constitutes an online harm.

Clauses 90 and 91 impose two duties on administrators.

First, they must not develop or maintain an online location in any manner that facilitates or permits online harm to take place – with the intention or knowledge that online harm is likely to take place. This duty covers administrators that are complicit in the online harms, and Members might be aware of the infamous chatgroup "SG Nasi Lemak" previously. This will cover the administrators of such a chatgroup.

Second, when notified of harm, they must act reasonably – more specifically, they must take reasonable care to assess if there is harm and if so, to take reasonable steps to address it. Clause 94 therefore imposes a similar duty on platforms to act reasonably when notified of harm.

I want to make clear to Members that the duty to take "reasonable steps" does not require the platform to do constant surveillance and monitoring. Their liability arises only when an actor fails to act reasonably after receiving proper notice. They are not liable if, through no fault of their own, they did not receive the notice sent by the victim.

In assessing reasonableness, the Court will consider the circumstances of the case, including the seriousness as well as the persistence of the harm.

Let me illustrate this: an administrator or platform that receives notice of a harmful post for the first time may act reasonably by simply removing that post if that is the appropriate remedy under the framework. But if the same account repeatedly causes harm in the same way, simply taking down each time the post is put up may no longer suffice. In such cases, taking stronger action, such as suspending or disabling the offending account, may be the reasonable steps required under clauses 91 and 94 and that is what I meant when I said you assess the entire factual matrix and situation holistically.

What is “reasonable” will therefore depend on the facts. The Courts can take into account factors such as the nature of the conduct; the context in which it occurred; and the effect and impact on the victim.

For example, putting up a post or creating a website to whistle-blow on serious misconduct may well be "reasonable" if done for a legitimate purpose and in a proportionate manner, even if it might cause harassment, or might be considered as online instigation.

There are also safeguards to address concerns that administrators and platforms may be inundated with frivolous notices or notices with insufficient information. The Bill provides that the particulars, which an online harm notice must contain are to be prescribed. We set it out clearly in a prescribed form so that the categories of information is known upfront and this ensures that only genuine, properly documented cases trigger the duty to act, and the administrators and platforms have enough information to take "reasonable steps".

Ultimately, it is the Court that will look at the facts of each case, weigh the totality of the evidence and decide whether a claim is made out and if so, what remedy should follow and against which party. These are all fact-sensitive judgments that reflect the diverse realities of online interaction. The Bill therefore avoids the use of fixed or rigid formulas, to allow the Court to develop the law incrementally, while at the same time, keeping the focus squarely on online safety and responsibility.

The Bill recognises that platforms and administrators need not proactively scan for all harms. They only have to act responsibly once notified. Take together, these duties encourage vigilance without imposing impossible burdens. They reflect a simple idea embodied in many legal principles, which is, if you control the space, then you must play your part in keeping it safe.

Next, Sir, let me turn to remedies. If a victim successfully establishes a claim, the victim must have access to effective and fair remedies. Under clause 96, victims may seek damages that the Court finds just and equitable, and other heads of damages that the Minister may prescribe in regulations, such as compensation for loss of earnings or an account of profits where perpetrators benefited from the harm.

The intent is to ensure that victims are properly compensated and that wrongdoers are not allowed to benefit from their behaviour. For some harms, therefore, the victim's earning capacity or livelihood may be affected and they should be compensated for loss of future earnings or loss of earning capacity, as the case may be or as may be appropriate. In other cases, such as where intimate images have been put online and for sale and perpetrators profit from this harm, then, in those cases, an account of profits may be ordered so that the wrongdoer does not get to retain the benefits of the harm caused.

The regulations in this Bill reduce victims' uncertainty as to what remedies they are entitled to. But ultimately, it is for the Court to decide on the appropriate orders, based on the facts of each case.

Clause 98 introduces the concept of enhanced damages and empowers the Court to award such damages where a communicator or administrator persists with their conduct despite notice. We think that enhanced damages should apply to those who are the root cause of the harm, such as recalcitrant communicators or administrators who create harmful websites or chatgroups. We have, therefore, excluded the platforms. This framework is intended to incentivise and drive reasonable compliance and, in some cases, as quickly as possible. And the enhanced damages framework also compensates victims for any additional harm resulting from failure to comply. And we hope to drive overall a strong enough messaging with a deterrent impact on the actors in the online space.

Therefore, enhanced damages may be awarded to compensate the victim for additional harm caused by the refusal to stop the online harm, penalise the communicator or the administrator for bad conduct, and the Court will consider the overall justice of the situation, when assessing whether to impose enhanced damages.

In addition, clause 99 empowers the Court to issue injunctions, both interim as well as permanent, to stop harm swiftly. These injunctions operate independently of any direction from the OSC, giving victims complementary routes to relief. The OSC and the Courts operate independently of each other, and neither is bound by the decision of each other. The OSC seeks to act quickly and takes public interest into account in making its decisions. The Court decides any claim for statutory tort relief based on the applicable legal principles and a framework for remedies. Taken together, Sir, we believe that we have fashioned a suite of remedies that strike the right balance: accountability for wrongdoers; fair recourse for victims; and flexibility for the courts.

The statutory torts are designed to change and strengthen norms – to make self-responsibility a default in our online spaces.

And I come back to the time when the first motor-safety laws were introduced in the 1960s. There was an American publication, Automotive News, which lamented the passing of these laws with the headline, and I quote: "Tough safety law strips auto industry of freedom". There was fear; there was resistance in the industry.

But with the passage of time, history has proved those laws right. They made cars safer, saved countless lives, changed attitudes and mindset, and re-shaped how the industry designed every vehicle thereafter.

We hope that this Bill, with a clear framework, can also have the effect of setting the right tone for online behaviour, shape mindset and attitudes, of both users as well as service providers. They define what is acceptable and what is not. They will guide conduct not only through monetary damages, but through shared expectations made explicit.

Overall, our intention is that as our online norms mature, we will rely less and less on reports and lawsuits, because, like road safety, the law will have done what it set out to do, not just punish harm, but nurture the habits that prevent harm in the first place.

Sir, I move on now to address how the Bill handles anonymity. From time to time, anonymity can serve a good purpose – it allows users to speak freely, sometimes, obtain assistance and, on occasions, allows marginalised groups to speak up. But at the same time, it must not shield wrongdoing. And unfortunately, many online users abuse the power and privacy which online anonymity affords them.

I had earlier covered how online anonymity is a driver of online harms and leads to the Online Disinhibition Effect. Anonymity also exacerbates the impact of harm on victims.

First, victims may become more distrustful of those around them. They wonder who it is posting on their social media sites. They do not know if the perpetrator is, indeed, someone they might know. Second, victims will not be able to obtain legal recourse from perpetrators. By definition, they cannot commence legal proceedings, or enforce Court judgments, against an unknown person.

There are existing mechanisms currently available in Court, such as pre-action discovery and non-party discovery. All of these are mechanisms which can be used to obtain information about the identity of wrongdoers. But victims will still need to commence Court proceedings which may be costly and time-consuming. And so, we believe the Bill's proposed End-User Identification measures offer an accessible option.

To start with, clause 49 empowers the OSC to obtain information and documents for the discharge of its functions. This includes identity information of an end-user which is in the possession of platforms. This is akin to how law enforcement agencies are empowered to obtain such information for investigative purposes, akin but not similar.

Second, clause 52 empowers the OSC, where it reasonably suspects a user of committing an online harm, to require prescribed platforms to take reasonable steps to obtain specified information that may identify the user. This can take the form of the user's name or perhaps, verified phone numbers or credit card information, which can then be used to make further inquiries with telcos or banks.

This obligation to collect information is carefully scoped to target those users who are suspected of carrying out online harms. This is following close consultations that we had with industry partners who expressed difficulty with a general obligation upfront for platforms to collect information of all their Singapore users.

Third, clause 53 empowers the OSC to disclose the perpetrator's identity information to a victim or to their authorised representatives, upon receiving an application from the said victim. At the initial stage, disclosure will be limited to the purpose of enabling victims to bring their claim. We intend to eventually extend this for other purposes as well, such as allowing victims to safeguard themselves from the perpetrator, and to take proactive future measures.

Mr Speaker, we recognise that some may have concerns that these measures might intrude on users' privacy or go too far. Let me be clear: that is not the case, and we thought about this framework quite carefully.

The measures are aimed squarely at those who hide behind anonymity to cause online harm. They are not meant to affect ordinary users who act responsibly. In fact, for the vast majority and for most users, nothing will change. Most platforms today already require some form of verified contact or payment information at the point of registration.

Additionally, when the OSC discloses a perpetrator's identity to a victim, there will be safeguards to ensure that the information is protected and not misused. First, the OSC may impose strict conditions on how the information can be used, such as limiting the use of the information to seeking protection or pursuing legal remedies. Any breach of those conditions will be a criminal offence.

Second, the misuse of the information may itself attract legal consequences. For example, if the victim were to use the information obtained from the OSC to dox the perpetrator, that could, itself, be an offence under POHA or under an online harm under this Bill.

In short, the Bill has in-built safeguards. They balance and they protect both the victim's right to know as well as the perpetrator's right against misuse.

Sir, this Bill is designed above all, as Members can see from how I have articulated the framework and the schema of this Bill, to be as victim-centric as possible, to give swift accessible access to those who have suffered real harm online. And the provisions have been drafted with that goal in mind. Minister Josephine Teo spoke about how the OSC will be empowered to issue directions quickly to address harmful content, and Minister of State Rahayu explained the appeal mechanisms available.

Sir, the hon Member, Ms He Ting Ru has proposed two amendments which I would like to address. Her amendments speak to the removal of finality of an Appeal Committee's decision, and second, to add a right to appeal to the General Division of the High Court.

Sir, I would like the House to know that both the MinLaw and MDDI teams had carefully considered the appeal process, and it includes options similar to Ms He's proposals. However, we felt that we could not support them in this Bill and let me explain why.

These mechanisms will make the process slower, with less finality to the proceedings. It will make it more complex and ultimately, less accessible for victims.

Let me reiterate that the purpose of the OSC is to deliver speedy, practical relief to give redress to what has objectively been determined to be an online harm. Allowing repeated appeals would prolong litigation and each new appeal means fresh rounds of arguments, delay and also uncertainty in dealing with harmful content, as well as renewed anxiety for those already hurt, who quite likely will have to remain engaged throughout the appeal process. We expect that there will be likely higher case volume in OSRA cases, which also adds to the administrative load of the OSC.

And, Sir, the further point is this. If a case goes on appeal to the High Court, lawyers will probably be instructed. In such an instance, will there be equality in how this might play out? One can imagine – most platforms are very well resourced, and likewise, a number of administrators and content creators too. What happens when an individual victim might need to seek redress against one of these giants, with deep pockets, in Court and with lawyers? With the additional prospect of having to bear substantial costs in the litigation if one does not succeed? Overall, we fear that this will dissuade victims from coming forward. And over time, this will render the framework toothless, not because of the provisions, but because individual victims will find it more difficult to seek redress and might shy away. This will make the framework less inclusive, and we hope not to see that.

In contrast, we believe that the current framework already strikes the right balance. In the first place, the framework that you heard Minister of State Rahayu outline earlier, these are administrative decisions by the OSC, who assess the harm based on the prescribed factors in this framework and they make a poly-centric decision, taking into account policy and public interest considerations when deciding on whether it is a harm and if so, what the appropriate remedy ought to be. Such administrative decisions are subject to judicial review and not an appeal. In fact, this is not unusual.

Sir, at its heart, this Bill, as I said at the outset, is about empowering victims. The OSC's process is deliberately designed to be straightforward, fast and simple and focused on stopping harm quickly and hopefully, not spending time arguing about it. We think that Ms He's proposed amendments, though well-intentioned, would probably make that journey harder and not easier.

Mr Speaker, Sir, this Bill is pragmatic, proportionate and principled. It protects victims, sets fair expectations for online actors and strengthens trust and accountability in our digital commons. If we proceed steadily and work together – Government, industry and users – I believe we can keep our online spaces open, but also safe; vibrant, but also responsible.

Sir, we have shown clearly how online harms exact a cost – on individuals, on families and on the social fabric that holds us together. And as technology evolves, new harms will emerge. Our laws must, therefore, remain future-ready. We must be bold and innovative to stay ahead but also compassionate in how we protect those who are most vulnerable. This Bill gives victims a clear and practical framework to seek relief when harm occurs. It also sends a clear, unambiguous signal – that everyone who shapes our digital spaces in Singapore must act responsibly.

Through OSC, the Statutory Torts Framework as well as the End-User Identification provisions, we are building a coherent system of protection and accountability. Each prong complements the other: the OSC is a safety net, providing rapid relief to victims of online harms; the Statutory Torts Framework provides private remedies, and it also is the standard-setter, encouraging all actors to play their part; and the End-User Identification measures ensure that no one can cause harm from behind a mask.

Sir, public support for these measures has been strong across communities, professions and also generations. In the Public Consultation launched by MinLaw and MDDI in 2024, respondents expressed strong support for establishing a dedicated agency to address online harms – over 90% support; allowing victims to take legal action, such as seeking compensation in Court for private remedies on top of the OSC framework – over 95%; disclosing a perpetrator's user information to the victim for certain specified purposes – over 80%.

Sir, the Government started this work a long time ago. We started looking at developing this Bill as far back as 2021, even as the amendments to the Broadcasting Act and OCHA were being worked on. We spent close to five years carefully examining the issues. We conducted numerous surveys and studies into the issue of online harms in Singapore, the findings of which have been presented to this House earlier. We also partnered with the SMU Yong Pung How School of Law to organise the Online Harms Symposium, where distinguished speakers and panellists, including experts on online safety from around the world, shared their insight on key issues and solutions for online harms.

In addition to the Public Consultation exercise, we also met and consulted extensively with over 100 different stakeholders over the years. This includes local and foreign experts, foreign regulators, victims of online harms, social service agencies, lawyers and the Judiciary, the Ministry of Education (MOE) and other educational institutions.

We recognise that the impact of online harms may be felt and experienced differently across different communities. We have heard from various segments of society, including the youths, disability and the community groups. We learnt much from their experiences and their insights and their stories. We also conducted over 20 engagement sessions with technology companies in the past two years to ensure that the provisions in the Bill are robust, workable, feasible and can be carried out when OSC issues directions.

We discovered, through these extensive engagements, a shared belief that the online world should reflect the same values of respect, decency and fairness that we all know and often assume, and which guide us in the offline space.

Around the world, societies are grappling with similar challenges – in Europe, the United Kingdom (UK), Australia and the US. We are moving in step with these global efforts, but shaping our own path, our own course, contextualised and nuanced to what Singapore needs.

Sir, ultimately, law and regulation alone cannot keep our people safe. We will require a whole-of-society effort. Public education must teach users to protect themselves, and every user must take ownership of their safety and behaviour online.

But if we can do this together – build sound laws, responsible platforms and a thoughtful public – we will strengthen not only our digital safety, but over time, our social fabric. And in time, our online norms will not erode, but endure – grounded in respect, anchored in responsibility and guided by the same values that make Singapore strong.

Sir, that leaves me to express gratitude to a few persons and groups who have contributed deeply to our work. As mentioned by Minister of State Rahayu, we convened a Steering Committee who guided our team in shaping our policy. In particular, let me acknowledge two members from the private sector: Ms Stefanie Yuen Thio, joint managing partner of TSMP Law Corporation, member of the Sunlight Alliance for Action and founder and chairperson of SHE; as well as Assoc Prof Eugene Tan from SMU.

In addition, SMU Yong Pung How School of Law partnered us to organise our Online Harms Symposium. The sharing from the experts and survivors at the Symposium informed much of our thinking on this matter. SHE's surveys and research on online harms and their experience in running SHECARES@SCWO, Singapore's first support centre for targets of online harms, provided us with data and insights to refine our policy.

And, finally, all those, many from the public who responded to our public consultation or who have engaged with us with very constructive comments and suggestions or written to my Ministry to share their stories. Every story helps us to shape the contours of this Bill. We thank them for their suggestions over the years.

Mr Speaker, we have in the audience today, in the gallery above, a few who have contributed deeply to our work and gave valuable feedback in developing our proposals. We have members from SHE and SHECARES@SCWO – Natalie, Hemavalli, Lorraine, Saira and Si Han – who together with their teams, served as vital pillars of support for those experiencing online harms today.

We also have representatives from YouthTechSG – Ben, Zoe, Beatrice and Kok Thong – who, together with many others, shared the perspectives of young Singaporeans with us. We are deeply grateful for their partnership and commitment to making our digital spaces safer for all and we record our thanks and gratitude for the time that they have taken and the experiences that they have so generously shared.

Mr Speaker, Sir, I stand in support of this Bill, because for victims of online harm, speed matters and justice delayed is justice denied.

A deepfake now takes seconds to create. With powerful AI, harm scales fast and it is not rare. Surveys show that more than 84% of residents have encountered harmful content online and roughly one in three have faced harassment or bullying. In particular, as many of our Ministers and political officeholders have mentioned, for our youth, online and offline are one continuous reality. So, harm in one space, whether it is deepfakes, doxxing, non-consensual intimate content, bleeds into the other.

Yet removal through current Court processes can take months and even years and traces often persist after takedown. This is the gap where victims suffer. Mental health declines, some turn to self-harm, drop out of school or lose jobs. Embarrassments follow, harassment piles on and cancel culture amplifies the pain. These are real psychological injuries. Sometimes a mental prison is worse than a physical one and we cannot keep telling victims to wait another three more months for relief.

That is why, for years, I have consistently called for legislation to curb this danger. In the 2022 Penal Code debate, I urged for laws to prevent cancel culture from taking root. In the 2018 debate on online falsehoods, I argued that democracies depend on facts, transparency and truth in public discourse, and warned that social media algorithms turn some of the most sensational content viral and some of these contents are false and harmful.

That is why I welcome this Bill. It closes the gap and strengthens trust – the foundation of a "we first" society. It protects people, it does not police speech. Expression is not a licence to harass, and freedom comes with responsibilities. It gives our people, especially our youths, room to think and disagree without destruction – a healthy digital commons to engage, create, and learn without humiliation. It delivers timely justice because harm spreads fast and hours matter, and OSC can act in days to remove content, restrict accounts and disclose identities so victims can seek damages.

Sir, for online falsehoods, speed is justice. At the same time, I have a number of clarifying questions to the Ministers for this landmark Bill.

One, implication of the amendments added on top of this Bill. In the amendments put forward, the evidentiary threshold moves from "reason to suspect" to "reasonable grounds to believe". Also, the amendments called for a process to appeal to the General Division of the High Courts. And the amendments call for the inclusion of serious harms like abetment or encouragement of suicide and sexual grooming into this Bill.

But I believe these are already addressed in criminal law, including the Penal Code, sections 306 and 376E; and the Vulnerable Adults Act 2018. To me, these amendments seem to be a substantive change to the due process proposed by this Bill, with implications to the victims – more legal cost, delayed justice. As such, I would like to ask the Ministers to inform us of the implications of these additional amendments put before Parliament votes.

My second point is that, no wrong doors. We already have laws on harmful online content – POHA, Broadcasting Act, POFMA, the Foreign Interference (Countermeasures) Act and OCHA. If someone is harmed online, how do they figure out which law to use? Will OSRA overlap with these or fit in with them? Do victims have to report to several agencies or will the agencies coordinate? And if OSC and the Courts disagree about whether online harm has happened, what happens then? Who has the final say and how is the conflict resolved?

Three, mental health integration. Removing content is necessary, but not sufficient. Victims need healing, not just remedies. Protection and healing must move together. How will OSC work with community groups and appropriate services to support victims, especially youth, of online harms? Can we consider offering psychological support along legal relief? And how will OSC make sure that the staff are properly trained so that they can help without causing further harm?

Four, Personal Data Protection Act (PDPA) and disclosures. Victims need identity disclosure to pursue redress. Platforms must also, at the same time, meet data protection duties. Will regulations clarify how Disclosure Directions interact with PDPA obligations, including purpose limitation and legal‑obligation exceptions? And will good‑faith compliance with valid directions be recognised so platforms can act quickly without fear of breaching confidentiality?

Five, cross border and evasion. How will OSC enforce its directions outside Singapore and will claims under the statutory torts actually help in such cases? The Bill lets OSC tell platforms to collect certain user information, but a person can ignore the request, ditch old accounts and open new ones. In such cases, how will the Bill still get the perpetrator's user information?

My final point, six, preventing abuse and protecting identities. What guardrails will prevent the statutory torts being weaponised for "lawfare", by that I mean frivolous claims or claims without merits to harass or pressurise respondents? As the Bill empowers OSC to identify users and disclose that information to purported victims, what safeguards are there to prevent misuse of that disclosed data for reverse doxxing?

In closing, Speaker, Sir, this Bill makes our online space safer, especially for our youths. It strengthens trust, delivers speed as part of justice and holds platforms and perpetrators to account. With these clarifications, I support this landmark Bill.

Thank you, Mr Speaker, Sir. Mr Speaker, for many in my generation and those before, we remember life before the Internet when communication meant pagers and for some lucky few, tumbler handphones and the familiar sound of a dial-up modem. Some may recall an even earlier time, where snail mail is just known as mail.

But for today's generation, our children, students and younger workers, they have grown up knowing only a world that is integrated with online. For most of us, the digital space has become inseparable from daily life. It is where we learn, we create networks, we work, and forms part of our identity.

While the digital realm has brought immense good in connecting us, empowering creativity and expanding opportunity, it has also become a space where real harm can occur. Over the years, we have seen painful examples right here in Singapore that remind us that what happens online can wound deeply offline.

In one tragic case, a teenage student took her own life after enduring prolonged cyberbullying from classmates on social media. Hurtful messages and ridicule in the digital space compounded the pressures she faced in the real world.

There was also the incident of a National Serviceman sentenced to jail after uploading his ex-girlfriend's private images onto a website, tagging her name and social media handle. The images spread within hours and, despite takedown efforts, the victim continued to face humiliation long after.

There are many more examples of online harms and these are not isolated incidents. They show that online harms are not just virtual. They are real, they are lasting and they are deeply personal.

From the early days of Singapore's digital journey, we have always been mindful of the risks that come with greater connectivity and many have played their part in building a safer online space. At the ground level, numerous organisations have been promoting digital well-being and online safety education and many of my fellow People's Action Party (PAP) MPs have also done so in their constituencies. In Pasir Ris East, we began our Digital Well-being and Safety lessons in 2020 and later expanded them through the M³ Goes Digital programme to reach more members of the minority community.

At the legislative level, Singapore has taken progressive and steady steps to strengthen online safety. We enacted POHA in 2014, introduced the OCHA in 2023, and updated the Broadcasting Act and IMDA Codes of Practice to hold online platforms accountable for harmful content.

A safe and responsible digital environment has long been a key advocacy point among PAP MPs. In 2024, our Government Parliamentary Committee (GPC) for Communications and Information, then chaired by Ms Tin Pei Ling with Members Ms Hany Soh, Ms Jessica Tan, Mr Alex Yam and myself, filed a Motion titled "Building an Inclusive and Safe Digital Society." Together with more than 10 PAP MPs, we called for stronger safeguards against online harms such as cyberbullying and urged platforms to do more to protect users.

Hence, our MDDI GPC welcomes the Online Safety (Relief and Accountability) Bill, which builds on these foundations – from ground feedback, to action, to legislation and now continued advocacy in this House. This Bill strengthens the framework to ensure that victims of online harm can receive swift, effective and meaningful relief, while reinforcing the responsibilities of platforms and perpetrators alike.

Globally, jurisdictions such as the UK’s Online Safety Act, the EU’s Digital Services Act, Australia’s Online Safety Act, Canada’s proposed Online Harms Act and New Zealand’s Harmful Digital Communications Act are grappling with the same question: how to protect users while preserving legitimate expression?

Online harms evolve rapidly. To remain effective, our laws must first and foremost protect victims, provide quick and accessible recourse, for in the digital age, harm spreads faster than healing; and ensure that justice is fair, balanced and proportionate, shaping a safer and more respectable online space for all.

I therefore stand in support of this Bill, which takes a pragmatic and, as both Minister Teo and Minister Tong described as a "victim-centric approach," something which I agree. It focuses on timely relief, clear accountability and shared responsibility across all stakeholders in the digital ecosystem.

That said, I would like to raise three areas for clarification: firstly, the definition and application of OSRA; secondly, how we can provide better support for victims of online harm; and thirdly, the possible unintended effects of the Bill.

Mr Speaker, the Bill defines 13 categories of online harms, of which the first five – online harassment, doxxing, online stalking, intimate-image abuse and image-based child abuse – will be implemented first. Minister Teo shared that the remaining eight will be followed progressively.

The Bill defines online harassment as the communication of material that a reasonable person would consider threatening, abusive, insulting, sexual or indecent, and likely to cause harassment, alarm, distress, or humiliation to the victim. While this provides a sound foundation, I seek clarification on how the threshold will be determined for online harassment? Will it depend on the number of incidents, the severity of content, or the type of platform involved, something to which Minister of State Rahayu alluded to. How will the OSC determine what breached the online harassment threshold and how will we compare one versus the other?

I would also like to seek clarification about the timeline and criteria for introducing the remaining categories. Also, if a victim experiences a form of harm covered only in a later phase, will the OSC have discretion to act on the case?

For the law to be effective, the OSC's investigations and directives must be swift and timely. Hence, may I ask, what is the expected case load under the first five categories? In Handout 2, which was distributed earlier, on the USC User Journey, what is the expected time between "a file reported to OSC" to when the "OSC may issue direction against the communicator, administrator and platform?"

In the earlier speech, Minister of State Rahayu shared that victims can file a report after reporting to the online platform where the harm is occurring. How do we ensure that this proactive reporting is not abused?

Who will be the case workers? How will OSC case workers assess the incidents? Will they be doing it manually and how do we scale it up quickly? The key point I am trying to make is, how do we adequately resource the OSC with the expertise, the manpower and the technology required to ensure that victims can get effective and timely relief? We need to set the OSC up for success.

Lastly, the Bill empowers the OSC to issue directions to the communicator, administrator, or online service provider involved. Can the Government clarify what qualifies as an "online service provider"? Will this extend to closed group or platforms such as WhatsApp, Telegram and WeChat?

Many Singaporeans are administrators of private chatgroups among friends, colleagues, classmates or even strangers such as in Build-To-Order chatgroups. Some of the most hurtful cyberbullying can occur within such closed platforms. How will the Bill handle cases arising from private or semi-private groups, where the boundary between personal and public communication is slightly blurred?

Mr Speaker, while the Bill rightly focuses on timely and accessible relief, a truly victim-centric approach must not only enable victims to act against perpetrators but also help them recover from the harm. Victims may take civil action through a statutory tort and claim enhanced damages in serious or malicious cases, a clear message that accountability applies equally online. However, the Bill does not address psychological and social support, especially for victims of cyberbullying, doxxing, or image-based abuse. The emotional scars often outlast the digital content itself.

The OSC can also play a broader and more supportive role in caring for victims of online harm, acting as a one-stop point of contact that connects them to counselling services, community partners and social agencies that can provide the help they need.

In Handout 2, it is also noted that counselling support today is self-initiated by victims at the point when the harm is experienced. But we know that, for many, this is often the most difficult time to reach out, when the victims may feel alone, frightened, or unsure of what to do next. Can I suggest for the system to be a bit more proactive? So, when a victim files a report with the OSC, could there be an automatic prompt or offer for counselling support, so that help is not just available but actively extended? This small change could make a big difference, ensuring that victims receive structured, compassionate and timely support at the moment they need it most.

Working with the Ministry of Social and Family Development, Ministry of Education, the Institute of Mental Health and social service providers, the OSC could establish a Victim Support and Assurance Framework comprising a confidential helpline or online portal for guidance; access to counselling and trauma-support services; and collaboration to track and prevent recurring patterns of online abuse.

Being victim-centric means more than punishing offenders. The law must deliver justice but beyond that, it is must also help victims rebuild confidence and trust in our digital space. With such a framework, the OSC can be not just a regulator but a guardian of digital well-being for all Singaporeans.

Mr Speaker, while the statutory torts clarify duties of online actors, we must be alert to unintended effects. Clearer legal rights may lead to more frequent or premature litigation, as victims, intermediaries, or even platforms resort to the courts as a first response rather than a last. We must guard against creating a more litigious society, where disputes are too readily taken to court. This would only make it harder for victims, many who are already distressed, who may hesitate to come forward because of the financial, emotional and psychological toll of litigation, and the pain of having to relive their trauma. Some may also fear losing a civil case and being burdened with legal costs.

This concern is even more acute when victims face large, well-resourced companies or global platforms. Such entities could exploit the prohibitive cost of legal action to deter victims from pursuing justice. Many may simply give up, unable to sustain long or expensive proceedings.

Hence, I support the approach of limiting appeals, to ensure that relief remains timely and accessible. If every dispute were taken into lengthy court processes, many victims would refrain from seeking help altogether, something which Minister Tong has shared earlier too. Justice must never depend on one’s capacity to afford it.

That said, I urge the Government and the OSC to monitor the use of statutory torts closely, ensuring that they empower victims rather than intimidate them; promoting fairness, accountability and confidence in our online safety framework. Mediation and conciliation should be encouraged as first-line remedies, and the OSC could monitor the civil actions to ensure the law remains fair and accessible to all.

Ultimately, this Bill must empower victims, not overwhelm them, making justice attainable, not intimidating; strengthening accountability without fostering a culture of fear or excessive litigation. Mr Speaker, allow me to provide a summary in Malay please.

(In Malay): [Please refer to Vernacular Speech.] Mr Speaker, this Bill is extremely important because the digital world is now integral to our daily lives. Online spaces bring about many benefits for learning, working and networking, but it also comes with harms such as cyberbullying, sharing of private images without consent and doxxing, which can cause profound emotional impact.

All this while, the Government and our people have worked to protect online users. We already have POHA, OHCA and IMDA guidelines to ensure digital platforms behave more responsibly.

Many PAP MPs have also conducted programmes such as Digital Well-being and Safety, and at M3 Pasir Ris, we organised M3 Goes Digital to educate citizens about cybersecurity. This new Bill complements those efforts and provides greater protection to victims.

With that, I would like to seek three clarifications.

First, how do we determine online harassment and when will the remaining categories of online harms be implemented?

Second, the importance of support for victims, including counselling assistance and guidance that can be coordinated by the OSC.

Third, ensuring this law does not bring about an excessive litigation culture, especially against the bigger firms that have the ability to intimidate victims with high legal costs.

This law must ensure swift action and immediate relief to victims, so that justice can be achieved expeditiously. Our goal will be to build a safe and responsible digital space for all Singaporeans.

(In English): Mr Speaker, Singapore's online space must be a place of opportunity, not harm. The Online Safety (Relief and Accountability) Bill represents another important step in our collective effort to make our Internet a safer and more trusted place for all.

This Bill builds on more than a decade of sustained progress and advocacy from the ground-up initiatives by community partners, PAP MPs who have championed digital well-being and online safety, to key legislative milestones, such as POHA and OCHA, and many others.

Together, these efforts reflect our continued commitment to protect Singaporeans from online harm while promoting a culture of digital responsibility. The Bill before us today is not a standalone measure, but part of a long-term journey; one that combines education, legislation and compassion to safeguard every user, especially the most vulnerable.

This, Mr Speaker, is the spirit of our digital future – a connected society that is also a caring one where freedom of expression is matched by responsibility and where justice online is real and reachable as justice offline.

I stand in support of this Bill and the shared goal it represents: to build a safe, responsible and inclusive digital Singapore for all.

Mr Speaker, before I begin, I would like to first declare that I work for a financial technology company that may fall under the definition of online service provider as envisaged in this Bill.

Mr Speaker, the WP understands the motivations behind this Bill and we deeply wish to support it but we have some reservations.

Ms He Ting Ru has addressed the first question the Bill raises which is would it adequately protect the most vulnerable victims. She has outlined how we can strengthen that protection – I support these amendments.

I address the second question: will this Bill be fair, accountable and properly calibrated in its exercise of its powers?

The Bill grants the Commissioner significant authority to issue directions, compel removal of content, impose obligations on service providers to reduce engagement with material without the creator's knowledge. These are necessary powers to address real harms, but there are also powers that should be carefully designed for their intended purpose and appropriately constrained by institutional checks and balances.

I will address three amendments that address our concerns about the Bill's architecture and six areas requiring Ministerial clarification.

Let me begin with our proposed amendments.

First is raising the threshold for state action. Clause 26 of the Bill sets the threshold at which the Commissioner may issue directions. The current text reads "reason to suspect". We believe that threshold is too low, too low for the powers being granted. "Reason to suspect" is a subjective test, it permits actions based on intuition or preliminary information without requiring objective evidence.

Our proposed amendment raises this to "reasonable grounds to believe". This is the same standard used in comparable legislation overseas, including the UK's Online Safety Act and Canada's proposed Bill C63, or their Online Harms Act. This requires evidence that would satisfy a reasonable person, not merely a suspicion. Some may argue this is semantics. We do not believe this is so.

In the UK Act, they explicitly distinguish between "reasonable grounds to suspect" in order to begin an investigation as opposed to "reasonable grounds to believe" to take enforcement action. We believe that this is the right approach. The former permits inquiry, the latter permits coercion. The distinction matters. If we are serious about protecting victims, we must be equally serious about ensuring the Commissioner's enforcement powers rest on evidence, not suspicion. Our amendment achieves both.

The second set of amendments I will be addressing concern legitimate discourse.

Clauses 9, 11 and 19 define three online harms: online harassment, non-consensual disclosure of private information and instigation of disproportionate harm. These definitions are necessary, but we think they are incomplete.

Let me give three scenarios.

First, a citizen posts fair criticism of a public official's conduct. If a reasonable person were to conclude that that criticism is, "abusive or insulting", and it is likely to cause the official distress, under clause 9, this could be harassment. Secondly, a victim of harassment publishes text messages from their harasser online as a call for help or perhaps, a warning to others, under clause 11 this could constitute non-consensual disclosure of private information. In the third scenario, a journalist publishes leaked documents exposing corruption in a government-linked entity, but under clause 11 this could also fall under the definition of non-consensual disclosure.

Mr Speaker, I do not suggest these outcomes are intended, but the Bill, as drafted, seems to permit them.

The definitions contain minimal carve-outs. There are no obvious exclusions for public interest and they do not go far enough in recognising that not all disclosures of private information are harmful and not all uncomfortable speech is harassment.

So, our amendments insert these safeguards.

For clause 9 we add that communication is not harassment if it constitutes fair comment on a matter of public interest. This is drawn from the established common law defence to the tort of defamation.

For clause 11, we add that disclosure does not fall within the definition of if the public interest in disclosure outweighs the public interest in privacy. We list seven examples – including exposing wrongdoing, informing the public on matters of significant concern and protecting public health and safety. This amendment is modelled on a well-established balancing test in comparable common law jurisdictions, including the UK. As drafted, it is also very similar to the test as codified in the Australian Privacy Act.

For clause 19, we add communication does not constitute instigation if it relates to a matter of public interest.

Our amendments do not weaken the Bill. They sharpen it. They ensure that the Commissioner's powers are used to protect victims, but not to chill legitimate speech. They prevent this Bill from inadvertently silencing criticism, investigative journalism or public interest disclosures.

The third set of amendments that I will be addressing establishes independent oversight.

The Bill establishes an interim appeal mechanism. Clause 60 creates an appeal panel whose members are appointed by the Minister. Clause 63 provides the right to appeal Commissioner decisions to an Appeal Committee drawn from this panel. Crucially, clauses 63(5) and (6) state that no further appeals will be permitted beyond this Appeal Committee. This makes a Ministerially-appointed committee the final arbiter.

Our amendment deletes these two clauses and sub-clauses and inserts a new clause C, establishing a right of appeal to the General Division of the High Court. Our proposed appeal mechanism is not unlimited. It is confined to three grounds, a point of law; secondly, that the harmful activity did not occur; and thirdly, that compliance is not technically feasible.

This is similar to the appeal mechanism in POFMA. This ensures that the Courts do not become a general review body for every single Commissioner's decision but remain available as an independent check on questions of legality, fact and feasibility.

Mr Speaker, this is not about distrusting the Commissioner or the Minister. This is about institutional design. When the state exercises coercive powers, especially a power that could affect livelihoods, reputations and businesses, there must be a route to independent judicial appeal of cases.

So, some may say that our proposed amendments may introduce additional burdens on the Courts.

I also appreciate Minister of State Rahayu and Minister Tong's earlier clarifications as to the policy considerations at play here. Minister of State Rahayu referred to the inherent right of judicial review, which is still available as there is no ouster clause within the Bill. We appreciate that judicial review is always available, but the scope is generally limited, in this case, it will be generally limited to the process of the Commissioner making their decisions and not about the merits of the case itself.

The proposed amendment we have tabled is very limited in scope. It proposes expanding this right of appeal to the Courts to three limited grounds. We understand that a balance needs to be struck and we are striving to achieve that.

With reference to Minister Tong's suggestions that that there may be a David and Goliath situation if we avail of appeals to the Courts. We believe that the existing provisions within the Bill, as drafted, notably section 63(4) which provides that there is no automatic stay on directions, even when an appeal process is proceeding, helps to ameliorate that concern. There will be no continuing harm to victims as the appeals process is proceeding. Furthermore, when we avail of the Courts, there are mechanisms that can address some of these concerns as well, such as in-camera private proceedings as well as gag orders to protect the identities of victims.

Mr Speaker, beyond these three amendments, I now turn to six areas where the Bill requires clarifications from the Minister.

First, I would like to address the scope of doxxing. Clause 10 defines doxxing as publishing identity information where a reasonable person would conclude was likely to have been intended to cause harassment, alarm, distress or humiliation.

Here is my concern. If a victim identifies their hitherto anonymous harasser online to warn others in the community, could a reasonable person conclude that this was intended to cause the harasser distress or alarm? The Bill does not answer this; so, I ask the Minister, will victims who expose their harassers in this manner be caught by clause 10? If not, what safeguards exist in the Bill's design to prevent this?

Second, I would like to seek clarifications on standing to appeal where directions are given to platforms. Clause 28 sets out who receives Part 5 directions. Some, including the stop communication and restraining directions, can be issued directly to the communicator, but others, such as Access Disabling, Account Restriction and Engagement Reduction Directions are issued to platforms or administrators. When a direction is not issued to the communicator directly, can they still appeal?

Clause 61(1)(e) allows the recipient to appeal, but the communicator is not the recipient. The subsequent subsection says that they may appeal only if they fall within a description, the Minister may prescribe under clause 82. So, I ask the Minister, will the regulations be made to ensure that communicators have standing to appeal directions that restrict their content, even if those directions were not issued directly to them or will this remain subject to Ministerial discretion?

Third, the meaning of prescribed connection to Singapore. I understand Minister of State Rahayu has addressed this example. She has given the example of long-term residents in Singapore who will fall under the ambit of this provision. We would like to seek further clarifications.

So, it is clear that Citizens, Permanent Residents (PRs) and long-term residents are eligible to make a report. We would like to understand what else the ambit of prescribed connection in Singapore could mean. Does it cover a foreign spouse on a long-term visit pass harassed by someone in Singapore? Subsequently, does it also cover a former Singaporean resident who has now moved overseas, who is still being targeted by the Singapore-based individual? So, I ask the Minister who is protected by this Bill, who is excluded and whether the prescribed connection will be made clearer?

Fourth, the exemption for public agencies. Clause 4(2) states that public agencies cannot be given directions or orders under Part 5. Part 5 contains the Commissioner's enforcement powers. This means that if harmful content originates from, is hosted by or is facilitated by a public agency, the Commissioner cannot compel the agency to act. Clause 4(3) goes further. Public agencies cannot be sued under the civil proceedings – provisions in Parts 10, 11 and 12.

So, Mr Speaker, online harm is online harm, regardless of its source. A citizen harassed through content on a Government platform or by a Government account experiences the same distress as one harassed on a private platform or by a private account. So, I ask the Minister, why are public agencies exempt from the Commissioner's enforcement powers and from civil liability? What is the policy rationale for the asymmetry and what recourse does an individual have if they experience harassment, perhaps, from a rogue public employee using an official account?

Fifth, Engagement Reduction Directions and Class of Material Directions. Clause 40 grants the Commissioner power to issue engagement Reduction Directions. This allows the Commissioner to require a service provider to reduce the engagement of end users with a Class of Material Direction without removing it. Clause 41(3) explicitly states that it is not necessary to give any person who may be subject to a Part 5 direction, an opportunity to be heard before the direction is given. This means content posters are not necessarily informed, the content stays online. It remains visible to the poster; it also remains visible to the victim. None will be aware that the contents' reach has been throttled. The victim will continue to see the harm, wondering if the report achieved anything at all.

Mr Speaker, this is a shadow ban. So, my question to the Minister is simple: what is the envisaged use of this particular direction? Surely, in cases of legitimate serious harms, one of the other more forceful directions is the better solution for victims, and if this is meant as a less forceful direction, why does the Commissioner see the need to intervene at all?

Clause 40, along with clauses 30 and 33, is also a Class of Material Direction, so this raises related concerns. These Class of Material Directions allow the Commissioner to issue Stop Communication Directions against entire categories of content, identified by specific identifiers, such as a username, a term, an online location.

There is a high risk that this will function as a digital dragnet that may capture legitimate content alongside harmful material. I give an example. If a victim's advocacy group reporting on an emerging category of online harms, as a warning to the community, posts about it, they may see that their material is swept up by mistake.

So, on Class of Material Directions, I have two questions: what safeguards exist to prevent over-blocking and what remedy exists for users whose legitimate content is caught inadvertently by the directions?

Sixth, and finally, I address consistency in decision making. Parts 10 to 12 of this Bill establish statutory torts that will be adjudicated by the Courts. These will generate judicial precedence and publish decisions. There will be transparency and consistency through the common law process.

But Part 5 directions, which are quasi-judicial in nature, there is no requirement for publishing reasons or directions. There is no public record of how Commissioner interprets and applies the definitions of harm in Part 3 and other sections. So, I ask the Minister, will the Commissioner be bound by prior decisions that they have made, or will each case be decided on fresh discretion? And secondly, will the Commissioner be bound by clarifications given today, in this House? If the Minister states that, for example, Clause 10's doxxing definition does not capture victims exposing their harassers, can future Commissioners be held to this clarification?

Mr Speaker, this Bill introduces fundamental protections for victims of online harm. We support these protections. Our amendments ensure that the powers are exercised on evidence with safeguards for legitimate speech and with independent judicial oversight, but oversight is meaningless if those affected cannot assess it. So, the Bill's ambiguities on doxxing, on standing to appeal, on coverage, on public agency exemptions, on shadow bans, on consistency in decision making, these are not minor. They go to the heart of who is protected, who is excluded and whether the regime operates fairly. I hope that in the course of this debate, that we can get clarity and comfort on these questions.

So, I urge the Minister to accept our amendments. If the Minister declines, I ask for clear answers to the questions I have raised. I think Singaporeans deserve a regime that protects victims and respects fairness in equal measure.

Mr Speaker, Sir, I rise in support of the Online Safety (Relief and Accountability) Bill, because behind every clause of this Bill is a face, a family and a cry for help. This is not just a Bill about digital platforms or statutory duties. It is about people, especially our young, our parents and those whose lives have been shaken by what is said or spread online.

Mr Speaker, the Internet moves faster than our feelings can heal and, indeed, faster than our reputations can be rebuilt. A harmful post, an exposed address, a doctored photo – once uploaded, it spreads in seconds. That is why the speed of takedown must match the speed of harm.

In an earlier debate, I shared the story of a young lady, Ms K, living in Punggol who was doxxed by her best friend's ex. He had created an account, just for doxxing and had put up one post with her face, name, place of work and where she used to study. When she came looking to me for help, the post had been up for two weeks. But even with the help of her loved ones, friends and, indeed, the extended Punggol family reporting the doxxing, the post was still not taken down even after a month had passed. I reported this as well, but was told, two weeks after I reported, that the platform is experiencing high volumes of reports and that I should report on the post, and not the account.

Note, this was a one-post account. It finally took several interventions by Government agencies and a certain trusted reporter/flagger NGOs to get the post taken down. By then, the post was already seen by a few hundred people, Ms K was aghast, but relieved and thankful. That is, until a few weeks later, she contacted me again, to tell me that her friend's ex, had doxxed her again, this time on the same account and on another platform. If this House feels weary just hearing about this, imagine how defeated she felt. This "Groundhog Day" is what many victims face – harm that repeats and hope that fades.

That is why I support the creation of the OSC, which is a one-stop centre to coordinate Government action and issue Takedown Directions across platforms. Victims should not have to file multiple reports to multiple agencies. OSC can help victims in navigating and simplifying the reporting process, from navigating which protective online harms statutes to invoke – is it POHA, POFMA, the Foreign Interference (Countermeasures) Act, OCHA, the Broadcasting Act – to ensuring that victims do not need to make multiple reports to multiple agencies and multiple social media platforms.

I also urge that Ministry to set clear service standards for takedown speed. This should not be measured not in days, but in hours. Online harm does not wait and neither should help. Because someone under siege, an hour can feel like forever. Families deserve both a response and a resolution.

I have several clarifications for the Minister: (a) how will the Commission ensure reports are simple to file, yet protected from false or vexatious submissions? In extreme cases, how do we guard against misuse of the system, and prevent it from becoming weaponised against the very same people it had set out to protect; (b) given that platforms themselves face high reporting volumes, how will OSC manage a potentially large caseload without compromising response speed; and (c) finally, while the first schedule covers harassment and doxxing, other categories, such as online impersonation and deepfake abuse, will only be implemented later. Could the Minister share the timelines and what recourse victims have, if their harm type is not yet covered or even identified?

Mr Speaker, I also support the Bill's recognition of the different players in the online ecosystem and, in particular, the responsibility of platforms. Platforms are not neutral bulletin boards. They are amplifiers of information, of emotion and, sometimes, of harm. When harm is done through them, they must be part of the solution, not silent observers.

I hope that with this Bill, delays in response will no longer be acceptable. Platforms must design systems that detect and throttle the spread of harmful content; provide real-time cooperation channels within OSC; and be transparent about how quickly and how often they act on Takedown Directions.

Families should not have to fight for accountability post by post. Responsibility must also rest with those who profit from the platforms that host them. And the Bill rightly requires platforms to "take reasonable measures" once notified of an online harm. But I would like to clarify on how this will be measured, what counts as "reasonable", and how OSC will assess compliance, particularly for major platforms with wide reach and significant influence.

Mr Speaker, laws can set boundaries, but they cannot mend a wounded heart. Real safety depends not just on legislation, but on the strength of our community: one that notices, listens and steps forward when someone is in distress. In Punggol Shore, we started a community initiative several years ago – the Punggol Positivity Pizza Movement. It began with our youth leaders and families who wanted to create a visible, warm symbol of encouragement for our young residents.

Their message was simple: "P.S. Punggol Shore – I love you. If you're going through a hard time, please reach out, because your community loves you."

Through the Positivity Pizza movement, volunteers from every corner of Punggol Shore – seniors, youths, women's groups, sports teams and religious organisations – have come together to crochet pizza plushie keychains, with a heartfelt goal: to make 20,000 pizzas, one for every young person living in Punggol Shore. Each slice carries a message of solidarity, "It's not okay to be not okay, reach out and your community has your back."

The inspiration for this movement came from a young resident, C, whom I met at a local mall. I had known her since her primary school days and when I asked how she was coping with secondary school, she confided that she was struggling with online bullying and thoughts of self-harm. As we spoke, she noticed a small keychain plushie on my bag and she said softly and shyly to me, "Ms Yeo, that is very cute. May I have it, so that I know that someone cared enough to stop to ask me if I was okay."

To C, if you are watching this, thank you. Know that you have started a movement that has touched thousands of lives.

Our Punggol Shore handcrafted pizza slices have become small but mighty ambassadors of empathy – connecting neighbours, bridging generations and providing that kindness, even if it is handmade and in real life, can still go viral. If cruelty can spread fast, then kindness must spread faster. When we see a friend being bullied or humiliated online, do not scroll past. Reach out, report, reassure. When we teach our children to stay safe online, let us also teach them to be kind online.

To complement this Bill, we, the Members of this House, the OSC, the community, must strengthen partnerships with our Family Service Centres, schools, youth networks and grassroots organisations, to build digital literacy, emotional resilience and first aid for the heart. So that our parents, our seniors and our young persons alike, can learn not only to protect themselves, but also how to support one another. Because real safety, Mr Speaker, does not come just from stronger laws. It must also come from stronger community bonds. Mr Speaker, in Mandarin.

(In Mandarin): [Please refer to Vernacular Speech.] In the online world, a harmful post or a malicious photo can spread across the entire island within minutes. So, we must act quickly. When harmful content appears, it must be taken down as soon as possible. Every hour of delay causes additional pain to affected families. We must also care for and protect one another. At the same time, I want to call upon all major online platforms to work together with us.

Let us join hands to create a safe, caring and responsible online environment. Harmful content is not someone else's problem, but an issue that affect every Singaporean family.

(In English): Mr Speaker, at the heart of this Bill lies our families and children who form the soul of our communities. Our online world must reflect the same values that we keep our physical neighbourhoods safe and that is, respect, responsibility and care. The Bill is part of this answer. Mr Speaker, I support the Bill.

Mr Speaker, I rise in support of the Bill. The Internet has transformed how we live, work and connect, driving innovation, enabling ideas and creating digital marketplaces where transactions occur at unprecedented speed. Yet, as with every invention, it has also created new opportunities for harm, for deception, harassment and abuse.

This Bill is therefore both necessary and timely. It equips vulnerable individuals with tools to protect themselves and ensures that our laws stay relevant in the digital age. It allows victims of online harm to have faster access to justice and we must ensure that in doing so it does not introduce new obstacles that make it harder for victims to seek redress.

Sir, allow me to highlight three key aspects that mark a significant step forward in addressing online harm in Singapore. The first, empowering victims through civil remedies; the second, introducing a new thought of online impersonation; the third and last, addressing inauthentic and manipulated media. Together, these measures strengthen accountability, protect dignity and promote responsibility in our digital spaces.

Sir, I am heartened that under Part 10, certain online harms, such as image-based child abuse and online impersonation are recognised not only as criminal wrongs but also as civil wrongs. This allows victims to seek redress directly without depending solely on state prosecution. This represents a crucial evolution. As the online sphere becomes central to daily life, individuals should have both the agency and responsibility to safeguard themselves – contributing to a safer shared online environment.

Mr Speaker, I particularly applaud the creation of the tort of online impersonation. By recognising this as a statutory civil wrong, the Bill offers victims a clear path to justice. They no longer need to wait for perpetrators to be unmasked or for criminal proof beyond reasonable doubt. Victims can act to restore their good name, remove false content and prevent further misuse of their identity. This reform affirms that digital identity is personal identity – restoring dignity and control to victims and signalling that accountability does not end at the keyboard.

Mr Speaker, I also support the introduction of the tort of inauthentic material abuse. With the rise of generative AI, manipulated images and fabricated videos are increasingly used to mislead the public, exploit reputations and erode trust. As technology blurs the line between truth and fabrication, our laws must evolve. Timely action against false and deceptive imagery will help preserve the integrity of public discourse and protect individuals from reputational harm.

Mr Speaker, but I would still like to seek clarification on a few points. First, on the Right-of-Reply User Directions under clause 34. This mechanism offers a fast, low-cost way for victims to correct false or defamatory content. However, will it complement or substitute existing remedies under the Defamation Act? Can a victim pursue both concurrently, or only if the Commissioner declines to issue a direction? Clarifying this interplay will help lawyers and users understand how best to obtain swift redress.

Second, on directions and orders under clause 26. While immediate intervention is sometimes essential, will the Commissioner conduct a preliminary screening before issuing directions? This ensures that powers are exercised not only quickly but proportionately.

Relatedly, would the Ministry consider issuing guidelines to clarify which harms will be prioritised for urgent attention? Such transparency would enhance trust and efficiency. This would also help manage public expectations, streamline workflows within the OSC and enhance confidence in its impartiality. Such a measure could serve as a more proactive alternative as compared to retrospective proposals like an annual report, which for example calls for the publication of annual reports and risk assessments for the preceding financial year.

Third, on the OSC. As the OSC is to be established by next year, I seek elaboration on its structure, its staffing and its expertise. Given its wide-ranging powers over privacy, speech and accountability, public confidence in its competence and neutrality is essential.

Mr Edwin Tong acknowledges that public education is essential, but I would also like to clarify if OSC is taking this educational role, promoting public awareness of online safety, informing users of their rights and responsibilities, and guiding victims to seek help because an informed public is our very first line of defence.

In conclusion, Mr Speaker, Sir, I fully support this Bill. It represents a comprehensive and forward-looking framework to make our digital space safer and more accountable. Globally, we see similar efforts: from the UK’s Online Safety Act 2023, Germany’s Network Enforcement Act and the EU’s Digital Services Act. Yet Singapore’s approach goes further by recognising new civil wrongs such as online impersonation and inauthentic material abuse, empowering victims beyond mere takedown or criminal models.

At the same time, I think we can also learn from others: proactive risk mitigation in the UK, systemic harm prevention in the EU, and clear removal timelines from Germany. These remind us that while the digital frontier is global, regulation can reflect and must reflect our local context.

I commend MDDI, MinLaw, the Attorney-General’s Chambers and all involved agencies for developing this landmark legislation. This Bill continues Singapore’s proactive response to online harms, giving victims real avenues for redress and ensuring that our digital space remains one of safety, trust and accountability. Mr Speaker, Sir, this Bill has my full support.

Mr Speaker, I support the amendments filed by Sengkang group representation constituency (GRC) MP, Ms He Ting Ru, and further elaborated upon by NCMPs Andre Low and Eileen Chong, and agree that these amendments strengthen the Bill for victims and better calibrate the significant powers Parliament is investing in the Commission.

I have some queries that cover about 10 clauses in the Bill as tabled by the Government for Second Reading.

First, clause 5(6) of the Bill allows the Minister to give the Commissioner directions of a general or special character, not inconsistent with the provisions of the Act. The Explanatory Statement does not shed light on what the ambit or scope of these two types of directions are. What general or special character directions does this clause envisage, which are not already fully captured or foreshadowed by the Bill? Can the Minister clarify what are some of the directions that the Minister envisages to give under this clause, or is this clause to be understood as a general one that confers broad and inchoate powers to the Minister?

Clause 8 empowers the Commissioner to issue advisory guidelines. As much as public attention on this Bill has focused on online harmful activities and the statutory torts that arise from them, I believe one important barometer for the success of the Bill will lie in its capacity for preventive education.

The Australian government's eSafety Online Commissioner's website is a useful guide to take note of as a one stop site to assist victims navigate the difficult terrain of online harms and educate young and old Singaporeans about the impact of their words and actions online, with a view to establish acceptable norms of online behaviour. To this end, can the Minister share the scope and outreach plans the Government has under this clause to address online harms, with particular focus on juveniles and young adults? Does it intend to copy the Australian approach with adjustments for the local context, or does it, for example, intend to work closely with the Ministry of Education to consider novel approaches? Are there specific autochthonous approaches on preventive education that pertain to online conduct and behaviours the Ministry plans on introducing suited for local circumstances?

Clause 11 on the non-consensual disclosure of private information is not clear as to what private information entails. While the clause leaves open clarity for future regulations, as drafted it reads too widely to be of practical application, and is thus unsatisfactory. On a plain reading, it begs more questions than it answers.

The clause states that private information is information about a person that is not widely available to the public at large. The Explanatory Statement to the Bill does not aid at all to clarify the ambit or context of "not widely available", and what it means. I hope the Minister can shed light on this clause and briefly outline how regulations will address the ambit of private information in this clause. If the private information is, for example, found behind a paywall of an exclusive and expensive online publication with limited circulation, would that information be construed as not being widely available?

Clause 20 covers a specific online harm, namely incitement of the enmity, which means the communication of online material that a reasonable person would conclude incites, or is likely to incite feelings of enmity, hatred or hostility against any group in Singapore. To this end, in finding the same, clause 20A of the Bill considers relevant factors, such as whether the material dehumanises any person or persons, or otherwise portrays them as less than human. For example, in the event a foreign embassy puts up a post that tends to incite, would the Commissioner issue an order against the embassy, an online service provider, or Singaporeans who share the post, or all three?

In connection to this, I seek clarity on the ambit of clause 22, which states that a victim of online harmful activity can be anyone who is a citizen or Permanent Resident of Singapore. Clause 22(1)C also includes an individual who has a prescribed connection to Singapore to make a report to the Commission. The interpretation section of the Bill does not define this prescribed connection, unlike how it defines prescribed online service provider.

Can the Minister clarify if the Bill can be expected to extend to all work pass holders in Singapore, including foreigners, or even diplomatic staff who make reports to the OSC seeking relief outside the excluded online harmful activities covering incitement of enmity and incitement of violence? And if not, what are the exceptions?

I move to Part 6 of the Bill covering information and end user identity matters. The Commissioner is empowered to require any person to provide in any form or manner any information or document, whether kept in Singapore or not. This includes requiring the person to provide an explanation of the information or document, including providing access by way of username, password or any other authentication information.

Clause 49(4) states that the Commissioner is entitled without payment to keep any information or document. Can the Minister confirm, notwithstanding Minister's comments about the usual police procedures, if anything in this part of the Bill allows the Commissioner to, for example, seize or retain a mobile device or laptop on grounds of an investigation into an alleged online harm under this Bill?

Clause 52 gives the Commissioner powers to obtain specified information about an end user and it does not matter if that end user is outside Singapore. I would like to enquire how the online service providers the Ministry has sought feedback from for the purposes of the Bill have reacted to this requirement and whether they would be able to comply, particularly online providers such as Telegram, an entity the Government is reported to have had difficulty with in the past. In this connection, can the Minister provide an update to the House on whether Meta has complied with the competent authorities' implementation directive to strengthen Facebook's measures against scams by 31 October this year as part of the OCHA. While I accept the relevant law is different, it would be useful for the House to understand how our laws interface with how promptly and the extent to which social media platforms can comply with them, and if not, why not?

Part 7 of the Bill covers reconsiderations and appeals. Clause 59(1)(d) requires the Commissioner to inform an applicant of the result of a reconsidered decision within a reasonable time. This clause does not require the Commissioner to provide reasons for his decision. In contrast, on the matter of the Appeal Committee in clause 64(5)(a), the Committee is required to state the reasons for his decision in respect of the appeal. Can the Minister share the rationale behind the treatment or the differentiated treatment, because from a layman's perspective, both are essentially appeals?

One approach to consider is for the Government or the Commissioner to provide reasons in both cases, or the Commission to provide reasons in both cases. Doing so would be helpful in reducing the number of appeals if the Commissioner's decisions are clearly explained at the outset, subject to the usual privacy concerns. Can the Minister explain the approach taken in the Bill so as not to give the impression that the first appeal to the Commissioner to reconsider the decision is not perfunctory?

Clause 94 creates the tort of failing to respond reasonably to an online harm notice. In this particular case, while it is foreseeable that the form of an online harm notice will be forthcoming, can the Minister shed light on what constitutes a reasonable response time from the online service provider? The illustration to this clause makes reference to the word "promptly" when a social media service responds to an online harm notice. This suggests that the interpretation of "reasonable" ought to be construed as "forthwith" or "almost at once". Would this be a correct reading and expectation of how quickly an online service provider is expected to react?

Sir, these are the clarifications I seek and I look forward to the Minister's reply. Notwithstanding the amendments in the name of He Ting Ru and subject to the clarifications from the Ministers, the WP will support the Bill.

Mr Deputy Speaker, the Online Safety (Relief and Accountability) Bill is timely. I have met with residents affected by online harms and Singaporeans have written to me asking for reforms. From how we stay in touch to how we work, shop, learn and govern, the Internet has evolved into more than just a means of communication – it has become a second life. And the divide between the online and offline worlds continues to blur as technology becomes increasingly interwoven into our daily lives. Our laws need to keep pace with modern day advances and the threats of today.

The recent report by SHE, "404 Help Not Found: Lived Experiences of Online Harms Survivors" sheds light on the harsh realities that define today’s digital landscape. Let me share the story of one of the online harms survivors mentioned in the report.

As a child, Zane experienced bullying in school. His peers took to online platforms weekly to spread negative remarks about him. Later in life, during his polytechnic years, Zane experienced doxxing. The perpetrator shared his name, school and course on Twitter and threatened Zane. These incidents had an impact on Zane in his growing years. He kept his pain to himself, fearing retaliation and judgement. The emotional toll eventually caught up with him. During his National Service days, Zane experienced panic attacks and had to seek professional help.

Zane's journey reflects how cumulative online harms can leave deep psychological scars; ones that linger long after the screen goes dark. And what happened to Zane is not rare. It is a glimpse into a culture where online harm has become casual, even accepted. Where cruelty hides behind screens, and victims often suffer in silence – many times too afraid, too tired and too ashamed to seek help.

It is important for us to be deliberate in shaping the tone of our online world for our children, our friends and our neighbours. Because the wounds may be invisible, but they are real – and they last.

The Bill is a step in the right direction. The OSC is vested with the authority to issue directions and orders to online service providers, communicators and administrators. Additionally, victims can turn to civil recourse and remedies for redress. In the age of AI, where false and harmful content can spread faster and cut deeper, such protections have never been more important.

The youths in particular are particularly and especially vulnerable to online harms. To find out what the youths thought of the Bill, I spoke with YouthTech SG, formerly known as Cyber Youth Singapore. YouthTech SG is a youth-led social movement and charity. One of their divisions is YouthTech Institute which focuses on helping young people navigate the risks and opportunities of technology, including online safety.

They wrote to me and expressed their support for the Bill, describing it as a signal for Singaporeans to foster a more cohesive, civil and wholesome digital environment. They also welcomed the Bill’s introduction of new avenues for youths to seek help and relief when harmed online; options that previously did not exist. While the Bill makes important strides in fostering a responsible digital environment, the online culture is ultimately one which requires users to build and that endeavour goes beyond legislation. Having reviewed the Bill, I find that the Bill is a meaningful step forward. It is unlikely to be a quick or perfect fix given how common online harms are and how deep rooted they are in our daily lives, but it signals our shared commitment to build a kinder, safer Singapore – both offline and online.

I support this Bill and would like to share a few reflections and seek some clarifications.

My first observation: the creation of the OSC and the scope of its enforcement powers is noteworthy. The speed of takedown is critical. While victims often suffer immediate harm, current means of redress are reportedly slow and sometimes ineffective. IMDA’s 2024 Online Safety Assessment Report found that major platforms acted appropriately only on half of the reported harmful content, often taking more than five days. The unfortunate reality is that victims are often left to navigate complex and sometimes opaque complaint processes, and structural challenges may further hinder platform responsiveness.

The OSC plays a vital role in addressing these gaps. Its powers, such as issuing Stop Communication or Restraining Directions, enable victims to take timely and decisive steps to safeguard themselves. Moreover, clause 26(1) provides that the Commissioner may issue a Part 5 direction if the OSC has reason to suspect that the online harmful activity was conducted. This allows OSC the flexibility to swiftly take down online harmful activity.

I see that Australia had similar experiences from which we can draw some lessons. The Australian eSafety Commissioner has the power to issue directives to address online harmful activities. Speaking at a 2023 symposium organised by MinLaw and the SMU Yong Pung How School of Law, the eSafety Commissioner from Australia highlighted the value of a dedicated and independent safety regulator in addressing online harms. In a Straits Times article, dated 26 September 2023, it was reported that she said: "Often when people come to us, they just want their images taken down, they don’t want to be tethered to their former partner through litigation, which, of course, can take a long time. So, what we provide is rapid assistance. I don’t think we would have the same degree of success with the platforms, particularly where there are grey areas, if we didn’t have these (remedial) powers."

To this extent, I note that there is a requirement for victims to first file a report to the online platform before approaching OSC. The duty imposed on the platforms also only starts when the report is accurately and properly filed to the platform. And therefore, the accessibility and ease of reporting on the platforms will be critical in ensuring access to justice for victims through OSC. I seek clarifications on how the Ministries plan to work with the platforms to enable this.

Secondly, I also seek clarification if the same content appears across different platforms and sites, how should the victim take action? Should the victim report all incidences he or she is aware of at the same time? Are OSC's directions specific? Or will we land in a situation where the victim is required to repeatedly go through the process of reporting to the platform and then to OSC for each piece of material? Can a victim apply to OSC for the pre-emptive ban of similar materials resurfacing?

Member Zhulkarnain had mentioned the advantages that OSC offers in being a more responsive and less adversarial mechanism as compared to traditional litigation. I will not repeat that, but I stand in support of his point.

My second observation is on anonymity. Victims of online harm may be reluctant to seek redress due to fear of repercussions, whether from the perpetrator, societal stigma, workplace consequences, or familial pressures. This hesitation is particularly pronounced in cases involving intimate image abuse, doxxing, or harassment, where the victim’s identity is often central to the harm itself.

I am heartened to see that the Ministry has been sensitive to this issue. In respect of court proceedings, the Bill envisages that the Rules of Court may be amended to empower the Court to order the redaction of identifying information, including parties' names, in court documents, to conduct proceedings in private, and to issue orders preventing the disclosure of parties’ or witnesses’ identities. I am concerned that in practice, however, victims may remain vulnerable and additional administrative steps may be required to secure their anonymity during the interim period before such orders are granted.

I wish to suggest for the Minister's consideration whether a more victim-centric approach would be to introduce automatic redaction, subject to judicial discretion to lift such redaction where appropriate. This would mirror existing protections under the Children and Young Persons Act, where proceedings involving minors are automatically anonymised, and the Women’s Charter, which provides automatic identity protection in the form of prohibition of publication or broadcast of information identifying victims or witnesses. This is applied under the Women's Charter in cases involving sexual offences. Extending automatic anonymity to victims of online harm, particularly those from vulnerable groups, would not only reduce procedural burden but also encourage victims to avail themselves to the options under this Bill.

Beyond the courtroom, I would be grateful for further clarity on how the Ministry intends to safeguard the confidentiality of victims during the reporting process. Specifically, how will the Ministry ensure that victims’ personal information remains protected, while also upholding the rights of alleged perpetrators to be informed of the nature of the allegations made against them?

My third observation relates to the new statutory tort of online instigation of disproportionate harm. Clause 87 of the Bill, which introduces the "tort of online instigation of disproportionate harm", which is designed to offer civil recourse to victims and their associates where public statements, made in response to a victim’s conduct, are likely to incite others to act in ways that could cause harm to the victim.

Civil action under this tort requires proof of actual or reasonably foreseeable loss or damage, while some heads of losses, such as financial losses, may be easier to prove, others, particularly psychological injury or emotional distress, have traditionally presented evidentiary challenges. I would be grateful for the Minister's clarification on how the latter cases will be treated.

My fourth observation is that the Bill seeks to address "inauthentic material abuse", which is any audio, visual or audiovisual material that has been digitally altered or generated that is false or misleading depiction of the victim’s words, actions or conduct, but is realistic enough that a reasonable person would believe that the material emanated from the victim. In this regard, a depiction of the victim saying or engaging in anything includes a depiction of the victim’s likeness saying or engaging in that thing.

The Bill would benefit from clarity on how likeness is assessed in the context of AI-generated content. For example, does resemblance alone suffice? Must there be an intent to impersonate or mislead? What evidentiary standards apply? In my view, these clarifications are especially important since resemblance may be subtle, stylised, or partial.

In closing, Mr Deputy Speaker, I welcome this Bill. It is a timely and necessary step towards preventing the propagation of online harmful activities. This is especially important today, when many of us, including young and old, spend much of our time online. The youths in particular are especially vulnerable to online harms, and I am heartened that this Bill provides additional avenues for their protection.

In particular, I am encouraged by the Ministries’ efforts to consult widely and engage with the youths while drafting this Bill. The youths form a significant demographic that would be affected and protected by this Bill. In their letter to me, YouthTech Institute shared with me that they participated in the public consultations and dialogues organised by the Government. They voiced their concerns and provided feedback on how the Bill can continue to stay true to its policy goals without chilling legitimate online participation by youths. They tell me that their feedback was seriously taken in and considered. By participating in the drafting process of this legislation, our youths felt seen and heard, inspired and empowered. The youths had this to say: "Thank you for showing us that policy can be made with us, not just for us." I am grateful to the Government for allowing the youth voices to be heard on a Bill that affects matters close to their hearts. I thank the Ministers and their teams for their efforts, and I am happy to support this Bill.

Mr Deputy Speaker, the objective of this Bill, if passed into law, is to enhance online safety for all in Singapore.

We operate every day in the digital world. It is part and parcel of our lives. While digital connectivity has brought undeniable benefits, a potent side effect has been the proliferation of online harms. Online harms are all pervasive, have the capacity to go viral very quickly and often bring immeasurable harm to the victims.

Victims find it difficult and challenging to seek recourse or relief from the mental and emotional toil that online harms inflict on them. For many, there is an innate helplessness in not knowing how to make it stop; how to make it go away. Ironically, even though it is the digital world, they have nowhere to run and nowhere to hide.

Many jurisdictions like us are also grappling with this phenomenon. Australia, New Zealand, Germany, the UK, the EU, the US and India have enacted varied legislation to address this.

In Singapore, the introduction of this Bill is, therefore, a timely one. I appreciate the policy intent of this Bill and the actionable levers that it seeks to introduce to address online harms in Singapore. I shall highlight three aspects that stand out for me.

One, this is a victim-centric Bill. Its main purpose is to provide quick and effective remedies and relief to victims of online harms, through a straightforward and accessible reporting platform, the Commissioner of Online Safety. All 13 classes of online harms are addressed from the victim's perspective.

Two, it addresses the pervasiveness of the online harm, by factoring in considerations, such as the number, frequency, nature and circumstances of the online communication. The Commissioner will take a holistic approach in considering each report to carry out his or her duties effectively and in good time.

Third and importantly, this Bill should not be seen as an attempt to regulate online content. Over time, it is hoped that this will promote accountable and responsible conduct that moves us towards a safer online environment in Singapore.

Mr Deputy Speaker, at this juncture, I would also like to address one aspect of the Notice of Amendments to the Bill filed by the hon Member Ms He Ting Ru.

I just like to make two observations to the inclusion of fair comment as an exception in certain clauses. Let us take clause 9 – online harassment – as an example. At the outset, I will say that I appreciate the good intent of the amendment. It is to carve out an exception of fair comment on a matter of public interest so that any such communication of online material will not constitute online harassment. Mr Andre Low, for example, gave an illustration of a public official: a person who made or gave a purported fair comment to a public official that could be perceived as an online harassment.

I will, however, make two observations. First, I note that the amendments do not propose a definition of what constitutes fair comment. As such, we have to look elsewhere for interpretation. One possible area is in existing defamation law where fair comment is a defence. There are four elements to the defence. But in simple terms, if the comment or opinion that a fair-minded or reasonable person could honestly made based on facts and importantly, it must be made without malice.

With that in mind, let us look at what constitutes online harassment in clause 9 of the Bill. It is the communication of online material that a reasonable person would conclude is threatening, abusive, insulting, sexual or indecent, and likely to cause a victim harassment, alarm, distress or humiliation. So, my first observation is this.

If we are to compare both definitions of online harassment and fair comment side by side, could a fair comment reasonably constitute as online harassment? Even with the example provided by Mr Andre Low, based on the definitions, is it possible for fair comment to be threatening, abusive or insulting? If the answer is no, then, the amendment, in my opinion, is unnecessary. In fact, I am of the respectful view that a fair comment should be a comment or opinion conveyed in a respectful, civil and courteous manner. I am sure such a comment will not be captured under clause 9 as online harassment.

This leads me then to my second observation. As I had shared earlier, my reading of the intent of the Bill is for the Commissioner to make an assessment that is first, victim-centric, how the online communication is received by the victim. Second, focused not on its content, but rather on the nature and manner in which the online communication is conveyed. And third, provide victims with quick and effective relief.

To necessitate the Commissioner to make further legal assessment on whether an online communication is fair comment, especially in light of my first observation will, in my opinion, distract from and potentially defeat the purpose of the Bill.

Having said that, Mr Deputy Speaker, while I rise in support of this Bill, I would like to take this opportunity to seek some clarifications on how the Government intends to address online harms suffered by a particularly vulnerable class of victims, children or minors.

In my maiden speech, I spoke about the concerns surrounding our children and their exposure to the digital space. Several Members of this House also raised similar concerns.

I had shared that, as a father of two Gen Alpha children, one of whom will be a teenager soon, no matter how much we try to regulate their interactions with the digital world, my wife and I worry. Where parents once told their children to beware of the strangers outside of the home, today, we must also guard against "strangers" and dangers lurking within, in the digital world. Even the home is no longer a safe sanctuary for our children.

Digital connectivity has dramatically altered the experiences of the growing up years. In the analogue world of the past, if you had an unpleasant experience at school or with your friends, you could go home, get "away" from the problem. You could, in essence, "disconnect". Today, these experiences follow you, everywhere. In fact, they become viral, multiplying and reinforcing the awfulness of the experience or encounter. It becomes the last thing you read or see before you sleep and the first thing you see or read when you wake.

For a child or a youth, this incessant barrage could lead to unthinkable devastating consequences. We have all read of such cases happening all over the world. I do not need to raise them here.

Mr Deputy Speaker, in this context, I raise two points for clarification.

First, how will the proposed Office of the Commissioner of Online Safety distinguish online harms between adults, children or youths? Second, whether the proposed Office of the Commissioner for Online Safety will take a differentiated approach in the reporting regime and the provision of remedies and reliefs when the victim is a child or a youth. I will elaborate on each of the above in turn.

Under the Bill, most of the classes of online harms depend on the "reasonable person" test. This is the test that the Commissioner will apply to assess and conclude whether the reported online material fulfils the criteria of an online harm. Currently, this is a single, uniform test to be applied to persons of all ages.

I accept that the Commissioner is within his or her remit to apply a differentiated standard of the "reasonable person" depending on whether the alleged victim is an adult or a child. This, however, as we can all appreciate, could be an extremely delicate exercise. There is, at present, little guidance provided in the Bill.

Other jurisdictions have drawn this distinction in their legislation. The Australian Online Safety Act 2021, for instance, draws a distinction between cyberbullying for children and cyber-abuse for adults. In fact, it appears that the Australian eSafety Commissioner has set up different channels in dealing with a range of different persons: adults, educators, parents, young people, children, seniors and women.

I believe that the perception of harm, such as what is threatening, abusive or insulting, between an adult and a child and even between minors of different ages is different. The type of online materials that may cause harassment, alarm, distress or humiliation to an adult as opposed to a child or youth, will also be different.

In fact, I thought Minister of State Rahayu had actually — when she shared Handout 4, if I can bring the House to Handout 4, paragraph 7 under (A) Online Harassment says this: "However, if the facts were different and B had disagreed with A, only by calling A's stance stupid and flawed, this is not necessarily online harassment. Criticism, even if strongly worded, is not online harassment if it does not involve what a reasonable person would conclude is threatening, abusive, insulting, sexual or indecent language, which is likely to cause harassment, alarm, distress or humiliation".

It is my humble opinion, Mr Deputy Speaker, that when words such as stupid and flawed are addressed to a child or a minor, it is perceived very differently. And so, when we look at the reasonable person test, we must be mindful that it must be viewed in the context of age and maturity.

Mr Deputy Speaker, this is a hypothetical, and I am certain that any future Commissioner will calibrate this accordingly. My only wish is to make the point that there needs to be a distinguishment between how we view online harms between adults and children. Studies have also shown that online harms impact children differently from adults. For example, online harms are likely to have a longer-lasting impact on children. Cyberbullied children experience depression, anxiety, suicidal ideation and self-esteem deficits that persist into adulthood.

Secondly, I hope that the Office of the Commissioner for Online Safety could consider a differentiated approach when setting up the reporting regime and the provision of remedies and reliefs when the victim is a child or a youth. Reports and surveys suggest that in the cases of cyberbullying of children or minors, under-reporting is prevalent. In February 2024, the then Ministry of Communications and Information released findings from their youth online gaming survey. The results are as follows: 17% or almost one in five gamers aged 13 to 18 felt that they have been bullied in video games by other players. Out of the 17% who felt bullied, nearly half or 48% did not take any action at all. Only about 8% spoke to their parents about the experience. Of the parents surveyed, one in four parents were not aware of who their child gamed with.

This is not unique to Singapore. Studies elsewhere have also reflected the under-reporting phenomenon. In another survey conducted by researchers from the University of California, Los Angeles (UCLA) on how 1,400 youths aged 12 to 17 deal with cyberbullying, 90% of the respondents reported that they did not tell an adult about the incident. The common reasons for this are the belief that they should "deal with it themselves", that is 50%; the fear of losing Internet access or face device restriction, or 31%. In other surveys, embarrassment, shame, fear of judgement or reprisal are also reasons children or youths cite for not reporting cyberbullying.

These surveys struck a chord with me. Ensuring that the reporting mechanism is one that provides confidence to our minors and a safe space to share their experiences and fears without judgement, is vital and crucial.

There are many reasons why a child or a youth may shy away from sharing cyberbullying incidents with parents or family. A compromising indecent image of themselves, for example. Will there be a reporting mechanism that would allow children or minors to directly report cyberbullying, even in the absence of a parent or a trusted adult? How will we reassure and encourage our children and youths that such a reporting mechanism is safe for their use?

To achieve this, it will likely entail a multi-pronged approach led by the Commissioner for Online Safety, involving the entire ecosystem that supports our children and youths from parents, educators, social workers, child psychologists and student guidance counsellors to name a few. In this context, I would ask the Minister of the Government’s views and plans to address the complexities and nuances of online harms suffered by children or youths in the digital world and, if there are plans to augment the Office of the Commissioner for Online Safety with the necessary support, such as a professionally trained team of experts to address these issues. Sir, notwithstanding my clarifications, I support the Bill.

Mr Speaker, I rise in strong support of this Bill. This Bill matters – to every parent who has ever worried about what their children see online, to every small business owner targeted by scams or false associations and to every young person who has faced abuse, ridicule and humiliation online, sometimes spilling offline. This Bill tells them something powerful: that the law sees their pain. That we are serious about safeguarding the dignity, privacy and well-being of every person online without stifling creativity, innovation, expression or enterprise.

This Bill is a major step forward. It is principled and practical – the product of years of careful consultation and iteration, including insights from last year’s Motion on Building a Safe and Inclusive Digital Safety in this House. In that debate, I called for an inclusive and safe digital society even in the age of AI. I said that platforms that profit from attention must also take responsibility for harm, including AI-generated content such as deepfakes.

This Bill delivers. For the first time, victims have a single front door and a clear pathway to redress. Perpetrators and platforms can no longer hide behind the anonymity of code or contract or algorithm. It holds intermediaries and platforms accountable when they fail to act reasonably upon notices. It clarifies the shared responsibility. We move beyond "blame the victim" or "blame the platform" to reasonable duties on everyone with power to prevent harm – users, group owners and administrators, and service providers and platforms, paired with liability for frivolous or abusive notices to prevent weaponisation.

It empowers good actors. By defining administrators’ duties, we give WhatsApp, Telegram, Discord group owners clarity: act reasonably and promptly when put on notice and you will be fine. Most community admins want safe spaces, the law should back them up with guidelines and safe harbours for good faith action.

It empowers the OSC to issue Right-of-Reply Directions, Reduce Engagement or amplification and award enhanced damages for persistent offenders – strong but calibrated safeguards for each type of online harm.

And it is forward looking. Even though the first phase focused on five online harms, the planned inclusion of inauthentic material or deepfake abuse and online impersonation recognises how AI is changing the very texture of online truth.

The Bill is also legally coherent. I have to admit that preparing to debate this Bill was quite a challenge, given the number of consequential changes to related laws such as harassments protections, remedies for identical reports, recognition of humiliation as harm. I commend the Ministry for the hard work aligning our justice system with the digital reality.

But I have three clarifications. First, for the new civil torts, what factors will courts consider? Speed of removal? Context? Labelling? Prior warnings? A short practice note would guide platforms, administrators and users, to act correctly and swiftly.

Second, timelines, including for appeals. What are the expected service levels from report to direction and typical reconsideration timelines? A fast track for high-risk harms would be reassuring.

Third, data protection and identity notices. When identity disclosure is required to pursue relief, what privacy safeguards and purpose restrictions will apply? We should deter abuse while protecting victims’ and bystanders’ data.

I also have five suggestions where we might go further.

First, moving from reactive to anticipatory protection. Today, duties are triggered after harm occurs, hence victim redress. Other jurisdictions now require major platforms to assess and mitigate certain risks before harm occurs, especially for children and vulnerable groups. A limited, risk-based set of anticipatory duties – not blanket surveillance but responsible design – would help to prevent harm.

Second, enhance algorithmic accountability. Platforms shape what we see. Algorithms are the new gatekeepers of truth. Platforms should explain, in principle, how their systems amplify or reduce harmful content. Much like how banks must disclose material risks, systemic accountability requires visibility.

Third, strengthen rehabilitation, digital literacy and AI awareness. Not every harmful act online is malicious. Sometimes it comes from ignorance, inexperience or immaturity. Courts should have discretion to combine compensation with digital-literacy or restorative programmes, especially for younger offenders. Long term safety is not just about punishment. It is about cultivating a culture of respect online.

Fourth, extend transparency to outcomes. Platforms could publish anonymised statistics: notices received, actions taken, average response times – so that the public can trust and policy-makers can measure progress.

Fifth, as technology evolves – so must the law. Periodic reviews are essential to ensure the framework remains effective, especially for emerging technologies like generative AI and immersive platforms.

Gen AI tools like ChatGPT can create harmful content. Left unchecked, this can put children and vulnerable users at risk. We should consider whether "content generating intermediaries" should be prescribed under our online safety framework with clear obligations on content moderation, age assurance and transparency. This could include, for example, documentation of model training or data governance, or disclosure of AI safety plans as has been passed in California.

Any adaption must also reflect the unique harm pathways, user prompts leading to AI outputs, in contrast to social medica and be risk-based to safeguard Singaporeans without stifling innovation.

Sir, I know these will be difficult issues and balances to grapple with. But in times or rapid technological change, our task is not to fear the future. Uncertainty should not make us pessimistic, it should make us better builders of systems that are inclusive, safe and resilient. Our laws must remain principles-anchored, technology-neutral and future-ready.

I have also studied the amendments proposed by the hon Member Ms He Ting Ru and other Members from the WP with interest. I appreciate the desire for safeguards against overreach, but as many have said today the law must put victims front and centre. A "fair comment" exception risks confusion and loopholes. Harassment is clearly defined, "fair comment", as seen by the debate earlier, is vague and legally complex – and may give bad actors an easy excuse.

Raising the threshold for OSC action from "reason to suspect" to "reasonable grounds to believe" would raise barriers for victims and delay timely relief. Waiting for victims to produce evidence of reasonable grounds to believe before acting is not protection. It is delay, it is harm, it is injustice.

The hon Member Mr Andre Low has made some comparisons to other jurisdictions – UK and Canada. We are not here to copy the UK and Canada. We are here to protect Singaporeans today.

But he also seems to believe the thresholds, the blanket threshold that applies for this instance to also to victim redress. And I actually wonder if the Minister can clarify that.

The UK Online Safety Act does use the reasonable grounds to believe in specific context, such as when defining the scope of regulated services, when defining the platform's duty of care, and for final enforcement, but I believe not to victim redress, which happens at the platform level, so it is not equivalent to the OSC stepping in to issue, takedown or redress orders. In practice, Ofcom's own guidance shows it can and does act on complaints and indications of harm. When Ofcom or the platform can act to prevent further harm in the early-stage assessment phase, while evidence is still being gathered.

That does not seem dissimilar to the balance that this Bill strikes. The "reason to suspect" threshold ensures victims are seen and helped immediately, while safeguards prevent abuse of the system, including appeals for reconsideration. This is practical, principled and humane lawmaking, putting people over procedure.

Can the Minister confirm my understanding?

Removing the finality clauses from the appeals provisions and adding a right to appeal OSC cases to the High Court could turn quick relief into long and costly litigation. The Bill already provides for a clear appeal architecture – reconsideration by the Commissioner, then an independent Online Safety Appeal Panel and Committee. This ensures that powers are reviewable, not arbitrary.

Let me make this relatable. Parents worried about cyberbullying, or professionals whose reputations are harmed online, do not care about legal theory. They care about timely, practical action – recognition of their harm and real, enforceable recourse and real accountability. And a system that prevents someone else from being hurt in the same way. This Bill gives them justice that is felt, not just written on paper.

Our goal must be quick relief and effective, accountability, not endless second guessing and legalese. And that is why I did not support these amendments.

Mr Deputy Speaker, "online harm" is not faceless. It is the woman who is a victim of doxxing; the honest small business owner whose reputation is destroyed overnight; the retiree who is targeted and manipulated online, losing his life savings; the woman whose private images become public forever; the young creator attacked for speaking up. But is also every Singaporean who has ever hesitated before posting something kind, for fear that cruelty will follow.

This Bill protects all of them – and reminds us all of our responsibility. Safety online cannot be legislated into existence; it must be built through culture, enforcement and example.

Some may ask: why does this matter so much for Singapore? Because the Internet is where our children learn, our businesses grow and our communities gather. When people feel safe online, they feel they belong. And when they feel they belong, they can speak, create and care for other another.

Our laws already protect people in physical spaces – at work, at school, in our homes better than most other places. This Bill extends the same promise into our digital spaces.

Mr Deputy Speaker, I recently met a teenage girl in Woodlands who was a victim of cyberbullying.  When I asked what she wanted most, she told me, "I just want to feel safe when I’m myself online." That, at its heart, is what this Bill is truly about. This Bill does not just respond to online harm. It rebuilds trust in the digital public space. It tells victims: "You are seen". It tells platforms: "You are accountable". It tells society: "We can be both open and safe".

So, I urge all Members today: let us pass this Bill with a united voice and continue refining it, so Singapore remains not only the safest place to live, but the safest place to be yourself, online and offline. Online safety is everyone’s right.

Mr Deputy Speaker, the Internet has become the space where many Singaporeans spend most of our time on any given day. It is where we work, socialise, play, engage services, amongst many other things.

Unfortunately, "online" has also become "where" and "how" malicious actors inflict harm on others. And the harm inflicted in the virtual world has far-reaching and deep implications in the physical world because they tend to permeate the very being of the individual victim and our communities, with potentially long-lasting damage.

A resident teenager came to my meet-the-people session recently. She shared with me the daily fear she lives under, being subjected to online harassment from a perpetrator based overseas, and who has been able to somehow gain access to her private online accounts to blackmail her and even threaten others, her friends in her social network. She yearns for effective enforcement action that can stop the perpetrator quickly and decisively.

Sir, this landmark Bill could not have come sooner for my resident and many others in Singapore. It reflects the Government’s commitment towards providing swift and decisive recourse to victims of online harms.

This is why I cannot agree with the amendments proposed by the WP to raise the threshold for the Commissioner of the OSC to give directions from "reason to suspect" to "reasonable grounds to believe". As Mr Andre Low says, "this is not mere semantics". The higher threshold proposed by the WP will mean that the Commissioner will need to take more time to put together the "reasonable grounds to believe" before giving directions, during which the victim continues to suffer in silence.

I agree with Mr Andre Low that we need institutional safeguards to ensure fairness for all parties, but I would argue these safeguards have already been provided for in the Bill. Clause 58 provides for the application for reconsideration of the Commissioner's decision. Clause 63 provides for the appeal against the Commissioner's reconsidered decision by the appeals board. And further, clause 69 makes the provision of false information an offence to be dealt with in the courts.

So, what these institutional safeguards mean is that the Commissioner can act swiftly to protect the purported victims while also injecting urgency on the part of the accused to clarify, seek reconsideration, make an appeal if they feel that the order was not the right one or if it was made with incomplete facts but it gives them urgency to do so and gives the victim justice.

So, given the rapid and exponential impact of online harms, this approach strikes the right balance.

However, this Bill alone is not a silver bullet to the threat of online harms, nothing can be given the complexity of it and we will require the enforcement capabilities, as well as local and international partnerships to implement the Bill effectively. So, I hope you can allow me to raise a few points and clarifications for the Minister.

First, given the severe mental and emotional distress online harms have caused, I would like to ask Minister whether the prescribed penalties of a fine of up to $20,000 and a jail term not exceeding 12 months for the failure to comply with OSC’s direction is adequate to serve as a deterrent to stop the offending actions promptly.

Also, is the Minister able to provide any guidance on how the courts should calibrate the penalties to be meted out? Because as a general principle, Parliamentary intent is key in the application of the law, and we should, as far as possible, make it clear for the purposes of administration of justice.

Next, I touch on the extra-territorial aspect of the offences under consideration. Clause 78 would empower our courts legally to deal with offences under the Bill which are committed outside Singapore, but the courts’ ability to deal with such offences would be contingent on the OSC’s efficacy in detecting the wrongdoing, getting the wrongdoers who are located overseas to comply with the directives and apprehending them if they do not.

Operationally and practically speaking, will the OSC’s arm of the law be long enough to reach the wrongdoers who may be located overseas? Who will be our key regional and international partners? Will the overseas Internet service providers and app distribution services comply with OSC’s order?

Relatedly, does the Bill contemplate online harms perpetuated by foreign actors, perhaps as a means of foreign interference?

Sir, I now move on to the issue of less conventional means of disseminating malicious content. Like the non-digital world, the Internet comprises dark and less overt corners as well.

So, while the Bill provides a comprehensive framework to deal with offences committed through the mainstream channels, are there sufficient and effective safeguards and capabilities against circulation by dark social media, dark traffic or perhaps even the dark web? For the membership or following of certain obscure forums can far outnumber groups on more familiar platforms such as Instagram or TikTok. Members of this House will know that SG Nasi Lemak Telegram chatgroup at its peak had more than 44,000 members.

The risk of involvement of organised or syndicated crime groups could further significantly enhance the scale of harm. Identity protection tools such as the Virtual Privacy Network (VPN) will add a layer of complexity to detection and investigation, enforcement efforts.

To be clear, I am not proposing the policing of private conversations. However, we ought to do our best to eliminate or reduce the gaps where they are.

Sir, my final point is on the resourcing of the OSC. Several Members have made the same point. Has MDDI formulated a projection of OSC's workload? As a to-be newly established agency, the OSC will have to deal with multiple issues on various fronts, besides the core work itself, including administrative matters, logistics, human resource and financing.

How will MDDI ensure that OSC will be endowed with sufficient resources to meet its core mission? Also, has MDDI considered if OSC’s officers will require enhanced powers, including possibly such as those under the Criminal Procedure Code to be more effective in its investigation and enforcement?

Sir, the OSC has a heavy and urgent task ahead of it. And in this starting phase, my view is that its focus should be on getting its effectiveness on the ground to stop cases of online harm. This is also why, while I understand WP's proposal for annual reporting to Parliament, I am not supportive of the proposed amendment to this effect at this point, given the administrative load it will add to the OSC in its current phase of development.

Mr Deputy Speaker, notwithstanding my clarifications, I stand in support of the Bill.

Mr Deputy Speaker, I rise in support of the Bill. The Internet is central to how we learn, we work and we connect. Yet the same space that empowers can also wound. Online harassment, doxxing and impersonation cause not only reputational harm and damage but also emotional and psychological harm.

In my years with working with students, I have seen how online words or even emojis can weigh heavily on a young mind. What begins as a passing comment or post can linger for weeks. It shapes how young people see themselves, how they trust others and how they engage in the world.

But we know that online harm affects beyond youths. A 2025 survey by MDDI found that one in three Singaporeans had faced online harm and two in five victims reported serious emotional distress. Behind every statistic is a person whose confidence, rest or peace of mind has been shaken. When digital interaction becomes constant and instantaneous; the pursuit to respond, to perform and stay visible has grown. You can ask many of the MPs here.

For many, that brings new forms of stress and exposure, the kind that erodes stress and confidence quietly but deeply. This Bill therefore is not merely about regulating conduct. It is about reinforcing dignity, trust and wellness and well-being in our digital lives. It recognises that feeling safe online is part of feeling secure in daily life.

Sir, allow me to touch on three key areas of the Bill: timely relief, accountability and collective protection and how each strengthens the well-being of our people.

First, through timely relief. The Bill establishes the OSC to give victims a faster and more direct path to help. Each day that harmful content remains online, it deepens harm. So, consider the Singapore Sports School deepfake case in 2024, where manipulated images of students were circulated. Counsellors reported anxiety and social withdrawal among those who were affected. A faster redress system could have reduced that distress and restored a sense of safety to our youths. This reflects a simple principle: we step in early, we support recovery and we prevent harm from deepening.

In the many community dialogues that I have held, parents often share that when their child is targeted online and they do not know where to run or turn to. To them, Police reports feel daunting, platform responses are slow, or may I say, very slow. The OSC will fill that gap with a trusted, accessible pathway for relief.

Second, on accountability. The Bill introduces statutory torts for online harms, such as harassment, stalking, doxxing and impersonation. It defines responsibilities for both perpetrators and for platforms. This is not about fuelling litigation. It is about setting boundaries. When expectations are clear, behaviour improves without constant enforcement. Ultimately, a safer online space grows from shared values from individuals, from families and institutions choosing responsibility over convenience.

I stressed, just as in my last Bill, platforms must do their part. For too long, victims have waited days for responses while hurtful content spreads unchecked. That is not neutrality. To me, that is neglect. These platforms hold immense influence over public life, and with influence comes duty. This Bill reminds them that their responsibility to people cannot end at their servers' edge.

The MDDI's mystery shopper tests found that more than half of valid reports of serious content were ignored or delayed. This Bill changes that, for platforms will now have a duty of care, not merely a policy preference. And when young people see that harmful actions carry consequences, lessons on digital citizenship gain meaning. It teaches them that civility is not weakness, it is strength.

Thirdly, through collective protection through representation. Sir, the third pillar extends protection beyond individuals to institutions. Authorised bodies, including trade unions and professional associations, can now represent victims.

This is significant because it extends workplace safety into the digital realm. Unions, such as the Healthcare Services Employees Union (HSEU) and the Singapore Teachers' Union (STU) can now act when members face online harassment in the course of duty. Over the years, we have seen so many examples and I am glad that this Bill comes in very timely. As a labour MP, I have met educators, healthcare workers and young professionals who face online criticism that turned pretty personal. They told me it is not just the comments, but the silence that followed, not knowing who to turn to and whether who would stand with them. That sense of isolation is something we must change. And I am glad this Bill enables the unions and institutions to act swiftly because the ability that can make the difference between despair and recovery is now here. When unions step in, they offer more than legal aid. They offer reassurance, solidarity and care – support that protects both moral and mental well-being.

In May 2025, the Ministry of Health reported that nearly two-thirds of community-care workers had witnessed or experienced abuse or harassment. Those numbers remind us that online harm is not abstract. It touches people who serve us daily with empathy and commitment. In my community work, I have also heard from residents and parents who say they feel exposed online, sometimes hesitant to speak up for fear that comments or images may be twisted or misused. That quiet fear undermines confidence. And this Bill offers a safety net, from institutional support to clear remedies and assurance that no one has to face such harm alone.

Sir, despite my strong support for this Bill, I have five clarifications for the Ministry to respond to.

Firstly, on service benchmarks and triage. Will the OSC set clear service benchmarks, for example, acknowledgement within hours and resolution within defined days, and include escalation lanes for frontline public officers and workers, such as healthcare workers or educators? A delayed response to these workers who serve us in the frontlines prolongs their distress and can affect their confidence in carrying out their duties.

Second, on parallel duties for platforms. How will the Government ensure that platforms remain first responders even after the Commission is activated, so that they cannot delay or defer action? For workers who engage publicly, like the teachers, nurses and the social workers, harmful posts can spread faster than formal action, and platforms must act quickly to protect those who serve the public.

My third point is on protection of privacy when identity disclosure is ordered. How will personal data be protected even as perpetrators are unmasked? Confidentiality is not about data, it is about psychological safety. Victims are more likely to seek help when they know their identity will be treated discreetly and with care.

My fourth point is on sectoral readiness and guidance. Given that schools, hospitals and community agencies are often the first to encounter online-related distress, will the relevant Ministries consider developing practical guidance to help staff activate these protections and respond with empathy? Such inter-agency readiness, even if outside of this Bill’s scope, will make the OSC's work more effective and ensure that victims receive consistent, trauma-informed support.

My last point is on union and professional-body activation. What will be the process for unions or professional associations to act swiftly on behalf of their members? Unions are often the first line of support when workers face online abuse. Clear guidelines and authorisations allow them to move quickly to offer legal protection, pastoral care and reassurance before harm escalates and I hope the Ministry can consider this. Mr Deputy Speaker, in Malay please.

(In Malay): [Please refer to Vernacular Speech.] Sir, this Bill carries three main purposes. It provides timely relief to victims of online harassment through the establishment of the OSC; it sets clear responsibilities for individuals and platforms so that all parties act more responsibly; and third, it allows unions like NTUC and professional associations to represent members who become victims so they do not stand alone when attacked in cyberspace.

Fundamentally, this Bill does not merely protect digital safety, it also safeguards our mental health and well-being, stability of the family and social cohesion. We want a society that has the courage not just to speak up, but to do so in a civilised manner, one that is active online but also mindful of the impact of words and actions.

Online safety is part of social safety. When citizens feel safe enough to interact politely and respectfully, both in cyberspace and the real world, we will strengthen trust, compassion and the sense of solidarity in our nation.

(In English): Mr Deputy Speaker, this Bill protects more than digital boundaries. It protects peace of mind. It complements our wider efforts in preventive health and mental well-being by reducing one of today’s quiet stressors: online hostility. When a teacher, a nurse or parent knows that the law stands behind them; when a student or a young person feels that their dignity will be defended; when the union can stand with its members in solidarity and assurance, then trust grows. Laws alone cannot guarantee kindness. But laws can make clear what we stand for: respect, responsibility and care. This is how we build a digital society that strengthens, not fractures, our people. Notwithstanding the clarifications I raised, I support the Bill.