AI policy updates: Singapore

Singapore launches a starter kit for testing AI agents for safety and reliability

IMDA released a starter kit for testing agentic systems, setting out a four-step approach (identify risks, test realistically without real-world harm, analyse results and logs, monitor and retest) and metrics such as tool-use accuracy and human-intervention rates. The AI Verify Foundation named its first five accredited AI testers the same day. Recorded from press coverage; IMDA's own page was not found.

Practical impact: Teams deploying AI agents can use it to plan pre-deployment tests and ongoing monitoring.

Update to the National AI Strategy released with 10 refreshed priorities

Milestone recorded for National Artificial Intelligence Strategy 2.0 (Singapore): Update to the National AI Strategy released with 10 refreshed priorities. Recorded from the official source listed.

Singapore launches a Model AI Governance Framework for Agentic AI

IMDA launched a voluntary governance framework for AI agents at the World Economic Forum, covering upfront risk bounding, human accountability, technical controls across the agent lifecycle and end-user responsibility, and said it was developing testing guidelines for agentic applications. Recorded from search extracts of IMDA and MDDI pages.

Practical impact: Organisations deploying AI agents can use it as a reference for governance and controls.

Singapore publishes the Model AI Governance Framework for Generative AI

IMDA and the AI Verify Foundation released a generative-AI companion to the Model AI Governance Framework with nine governance dimensions including accountability, incident reporting and content provenance.

Practical impact: Generative-AI developers and deployers in Singapore have a reference set of expected practices.

NAIS 2.0 launched by DPM Lawrence Wong

Milestone recorded for National Artificial Intelligence Strategy 2.0 (Singapore): NAIS 2.0 launched by DPM Lawrence Wong. Recorded from the official source listed.