Singapore Launches 'Nutrition Label' Guidelines for GenAI Chatbots and Tougher Personal Data Rules
Source: CNA
If you have ever wondered what a GenAI chatbot actually does with your data — or whether you can trust what it tells you — Singapore just gave you a single place to check. On Monday, the Infocomm Media Development Authority (IMDA) released two sets of guidelines that together represent Singapore's most...
If you have ever wondered what a GenAI chatbot actually does with your data — or whether you can trust what it tells you — Singapore just gave you a single place to check. On Monday, the Infocomm Media Development Authority (IMDA) released two sets of guidelines that together represent Singapore's most substantive move yet to govern how generative AI interacts with the public.
The first is a voluntary "nutrition label" framework for GenAI chatbots — infocards that disclose what the chatbot can and cannot do, how reliable and safe it is, how user data is handled, and how to report issues. Announced by Minister for Digital Development and Information Josephine Teo at the Singapore Data Festival, the guidelines encourage companies to consolidate information currently scattered across terms of service and privacy notices into a single, accessible format. "Like a nutrition label or medicine label, it gives users a single, consolidated place to find essential information about the chatbot they are using," IMDA said.
The second set of guidelines — issued under the Personal Data Protection Act — addresses how organisations should collect and use personal data for GenAI development. These advisory guidelines clarify several grey areas that have been a source of uncertainty for businesses. Organisations can scrape publicly accessible personal data without consent, but data behind paywalls or registration barriers requires a fresh assessment. Critically, companies must now specifically notify individuals when their data will be used for GenAI training — a broad statement about "product development" or "service personalisation" is no longer sufficient. They must also spell out what data will be used, what it will be used for, and how individuals can decline or withdraw consent.
The dual announcement signals that Singapore is taking a calibrated approach — voluntary guidelines for industry best practice paired with tighter enforcement of existing law for data protection. Several major organisations have already indicated they will adopt the chatbot infocard framework within the next six to 12 months: DBS, Google, Meta, OCBC, and Singapore Airlines on the private sector side, and the National Library Board and Health Promotion Board on the public sector side. The guidelines also break new ground by addressing the full AI value chain, clarifying roles and responsibilities for parties involved in developing, deploying, or supplying data for GenAI models.
Why it matters for Singapore: These two guidelines together close a growing gap between how fast GenAI is being adopted in Singapore and how well consumers understand what they are signing up for. The chatbot nutrition label gives users a practical tool — something between a terms-of-service document and a product safety warning — to make informed choices. Meanwhile, the PDPA clarification removes legal ambiguity for the hundreds of Singapore companies building GenAI features, giving them a clear compliance path while strengthening individual rights. As Minister Teo noted, many organisations already provide AI-specific notices today — the guidelines formalise what adequate notice actually means. For a city-state positioning itself as the region's trusted AI hub, this combined transparency-plus-accountability move sends a clear signal to both consumers and global investors.
