Singapore Launches Three Initiatives to Strengthen AI Safety and Resilience

Singapore's IMDA announced three initiatives at the AI Assurance Exchange: The Agentic Edition event to strengthen AI safety and resilience. The first is a starter kit for safety and reliability testing of agentic systems, providing a four-step process: identifying risks such as erroneous or unauthorized operations, building tests covering agent behavior, running tests and analyzing logs, and continuous monitoring and retesting, incorporating implementation cases from Cohere, KBTG, and others. The second is AI TAP, a certification program for third-party AI testing organizations launched by the AI Verify Foundation, with the first five organizations certified, valid for one year. The third is the Frontier AI Resilience and Safety Task Force (FAIRS), bringing together over 20 experts in telecommunications, media, cybersecurity, and AI, planning to release a report in the first half of 2027.