Regulatory Framework for Artificial Intelligence Governance in Singapore

Mr Speaker, Singapore supports the responsible development and deployment of artificial intelligence (AI), so that its benefits may be enjoyed in a trusted and safe manner.

We have introduced tools to help us achieve these goals. One example is the Model AI Governance Framework. Companies such as DBS, HSBC, Visa and Microsoft have adopted this framework to address key ethical and governance issues when deploying AI solutions. Another tool is AI Verify, a self-testing framework and toolkit to demonstrate responsible deployment of AI. It has attracted the interest of over 500 companies. [Please refer to "Clarification by Senior Minister of State for Communications and Information", Official Report, 21 April 2023, Vol 95, Issue 101, Correction By Written Statement section.]

The Personal Data Protection Commission will issue Advisory Guidelines on the Use of Personal Data in AI Systems under the Personal Data Protection Act within the year.

We engage our industry and international counterparts on AI-related issues through our Advisory Council on Ethical Use of AI and Data as well the Global Partnership on AI.

Similar to how we regulate data use, cybersecurity, misinformation and online harms, we will continue to review the state of the technology, our regulatory approach and its effectiveness, in order to uphold trust and safety in digital developments.

Thank you, Mr Speaker. Just a quick supplementary question. Given the rapid development of AI and the specific nature of AI which might be different from more traditional technologies, I was just wondering whether there was anything more particular or specific schemes or any protections to be put in place to protect citizens against AI-powered scams, for example, voice cloning, and whether these are actually being considered and what is the expected timeline for implementation.

Sir, I thank Ms He for her questions. I would take this opportunity to point out that in her questions is an assumption that AI is "one thing", and this is something that the public and commentators often treat it as.

AI is a group of technologies. Currently, there is a lot of excitement around ChatGPT, which is an example of a large language model based on machine learning techniques. When it comes to the issues of how we should deal with the impact of AI, its impact is going to be quite different in different domains and the impact is going to be quite different in how we regulate it, depending on the various components of what the AI tool rests on.

When the Member asks about what we will do to protect citizens from scams related to AI, it is not necessarily something that we need to be specific about AI regulation, it is something that we need to deal with in terms of our scam efforts, our initiatives around fraud and cybersecurity and criminal activity. And our colleagues in this House have spoken about that and we will continue to address that.

Similarly, when we talk about what we are doing around data, cybersecurity, online harms and misinformation, it is not that that is separate from what we are doing with AI, because those are things that AI-related tools may influence and may subsequently introduce harms to our public.

I just want to reassure the Member that we agree with the concerns, that we do have to address how technology is being used, make sure that the technology and the technological spaces are trusted and allows the development of initiatives that will benefit all of us and this includes protecting citizens from scams and efforts associated with that, whether AI-related or not.