What happened
On 2026-07-20, IMDA published voluntary guidelines setting a transparency standard for consumer-facing generative AI chatbots, calling for clearer disclosure of AI capabilities, limitations, and data-handling practices. Major firms (OCBC, DBS, Google, Meta) reportedly plan to use the guidelines as a reference. This is distinct from IMDA's earlier Model AI Governance Framework for Agentic AI (first published Jan 2026, updated May 2026) and from the PDPC's separate draft AI-training data-notification proposal (reported mid-July 2026) — confirmed via page metadata (datePublished=2026-07-20) but not yet cross-verified against IMDA's own primary publication page since only secondary press coverage was available in the search window.
Why it matters
Extends Singapore's voluntary AI-governance-by-guidance approach into consumer-facing disclosure requirements, adding a data-handling/transparency dimension to the existing agentic-AI risk-management framework. It signals continued national-regulator activity in the AI governance space that multinational tech/finance firms are voluntarily aligning to, which can create de facto industry norms ahead of any binding requirement.
Action needed
Firms operating consumer-facing GenAI chatbots serving Singapore users should review disclosure/UX practices against the new voluntary guidelines; track whether IMDA moves from voluntary to mandatory status.