Singapore has rolled out a new governance package for artificial intelligence dubbed MANGO, setting out deepfake rules and proposing an erasure right that would let people demand AI systems delete their personal data.
The framework is aimed squarely at the newest generation of agentic AI — systems that do not just answer questions but browse, book, buy and act on a user's behalf, raising harder questions about consent, accountability and how synthetic media is labelled.
The deepfake provisions target non-consensual and deceptive synthetic content, an area where Singapore has been tightening election and online-safety law. The erasure proposal mirrors the "right to be forgotten" in European privacy law but applies it to AI training and inference systems.
Singapore has built a reputation as a fast-moving test bed for AI governance, and its model frameworks are watched by regulators across Asia-Pacific. Tech firms operating in the city-state will be expected to show how their systems comply.
