Written Answer · 2026-08-05 · Parliament 15
Mandatory Labelling for AI-generated and Digitally Manipulated Content to Align with Code of Practice for Online Safety and GenAI Transparency Guidelines
Mr Alex Yeo asked the Minister for Digital Development and Information, in writing, whether the Ministry has assessed the case for regulation requiring indicative labelling of AI-generated or digitally manipulated content shown to Singapore users, similar to the transparency obligations in Article 50 of the EU AI Act, and if so how this would relate to the Code of Practice for Online Safety and the new voluntary GenAI chatbot transparency guidelines. Minister Josephine Teo replied that the Government regularly assesses its regulations and frameworks to ensure they remain relevant as technologies evolve. The Infocomm Media Development Authority's two Codes of Practice for Online Safety require designated social media and app distribution services to put in place systems and processes to mitigate Singapore users' exposure to harmful content, including AI-generated content. The Transparency Guidelines for Generative AI Chatbots encourage deployers to clearly explain to users the capabilities, limitations and safeguards of their chatbots. MDDI will continue to monitor international developments on labelling AI-generated or digitally manipulated content, keep track of the maturity of technical standards such as watermarking and digital provenance approaches, and assess whether they should be mandated. The reply made no commitment to a mandatory labelling regime and set no timeline.
Why it matters
Asked about EU AI Act Article 50-style mandatory labelling, MDDI still leans on IMDA's two Codes of Practice for Online Safety and will only assess mandating watermarking and digital provenance once those standards mature.
Key Points
- • The Member cited Article 50 of the EU AI Act in asking whether labelling of AI-generated or digitally manipulated content shown to Singapore users should be mandated
- • IMDA's two Codes of Practice for Online Safety require designated social media and app distribution services to mitigate users' exposure to harmful content, including AI-generated content
- • The Transparency Guidelines for Generative AI Chatbots are voluntary, encouraging deployers to explain chatbot capabilities, limitations and safeguards
- • MDDI will monitor international labelling developments and track the maturity of standards such as watermarking and digital provenance before assessing whether to mandate them
- • The reply commits to no mandatory labelling regime and no timeline
The Government's position is that existing instruments already cover the risks of AI-generated content — IMDA's two Codes of Practice for Online Safety govern harmful content from the platform side, and the chatbot transparency guidelines govern disclosure from the deployer side — so there is no rush to adopt EU-style mandatory labelling. Whether to mandate it will depend on international developments and the maturity of technical standards such as watermarking and digital provenance, with the Government reserving room to assess.
The questioner, Mr Alex Yeo, was concerned that Article 50 of the EU AI Act already requires transparency labelling of AI-generated and deepfake content, and asked whether Singapore should likewise legislate labelling of such content visible to local users, and how such rules would fit together with the existing Codes of Practice for Online Safety and the voluntary chatbot transparency guidelines.
Singapore is explicitly declining to follow the EU AI Act Article 50 route of legislated mandatory labelling, instead splitting AI-content governance across two existing tracks — platform obligations under the Codes of Practice and voluntary deployer disclosure under the transparency guidelines. By making "maturity of watermarking and digital provenance standards" the precondition for mandating, it hands the decision to technical standards rather than a legislative timetable, so no mandatory AI-content labelling law is likely in the near term.
"We will also keep track of the maturity of technical standards, such as watermarking and digital provenance approaches, and assess if they should be mandated."
Participants (2)
- Mr Alex Yeo
- Josephine Teo
Original Text (English)
SPRS Hansard · Fetched: 2026-09-04
36 Mr Alex Yeo asked the Minister for Digital Development and Information (a) whether the Ministry has assessed the case for regulation requiring the indicative labelling of AI-generated or digitally manipulated content shown to Singapore users, similar to the EU AI Act Article 50 transparency obligations; and (b) if so, how this will relate to the Code of Practice for Online Safety and new voluntary GenAI chatbot transparency guidelines.
Mrs Josephine Teo : The Government regularly assesses our regulations and frameworks to ensure that they remain relevant as technologies evolve.
The Infocomm Media Development Authority's two Codes of Practice for Online Safety require designated social media and app distribution services to put in place systems and processes to mitigate Singapore users' exposure to harmful content, including those generated by artificial intelligence (AI). The Transparency Guidelines for Generative AI Chatbots encourage deployers to clearly explain to users the capabilities, limitations and safeguards of their chatbots.
We will continue to monitor international developments on the labelling of AI-generated or digitally-manipulated content. We will also keep track of the maturity of technical standards, such as watermarking and digital provenance approaches, and assess if they should be mandated.