Singapore forms group to protect against AI cyber threats — Channel NewsAsia
Singapore has formed a task force of more than 20 experts and industry representatives to strengthen the protection of the infocommunications and media sectors against cyber threats related to or accelerated by artificial intelligence. As Channel NewsAsia reports, the Frontier AI Resilience and Security Taskforce (FAIRS) includes specialists in telecommunications, media, cybersecurity and AI.
Task force objectives
FAIRS will work in three areas: accelerating responses to threats amplified by AI; strengthening critical systems as artificial intelligence technologies become more effective at identifying vulnerabilities; and sharing information within the sector to detect emerging attack methods early.
The group is to publish a report with practical recommendations and best practices in the first half of 2027. Ng Chee Pong, chief executive of Singapore’s Infocomm Media Development Authority (IMDA), said that frontier AI models are changing the rules of the game in cybersecurity. According to him, some risks cannot be controlled solely by testing individual artificial intelligence applications.
In particular, FAIRS members will consider ways to strengthen and segment networks and systems so that individual vulnerabilities do not lead to large-scale disruptions. They will also explore opportunities to improve network monitoring and responses to threats that develop faster through the use of AI.
More current news is available on the UA.News Telegram channel Telegram.
Testing AI agents
At the same time, IMDA introduced the first version of a starter kit for testing the safety and reliability of agentic AI systems. Unlike conventional generative systems, such agents can act on behalf of users, use tools, access external systems and confidential information, and modify their environment.
The kit provides for four stages: identifying risks, including erroneous or unauthorized actions; developing realistic tests; analysing operational logs; and continuous monitoring and retesting. IMDA recommends assessing the agent’s entire workflow — its actions, access and compliance with permissions — rather than only the final result.
The AI Verify Foundation also named the first five companies accredited under the AI Tester Accreditation Programme: AIDX Tech, Knovel Engineering, PwC Singapore, Resaro and Vulcan. The programme is intended to help businesses find independent specialists to test the safety and reliability of generative AI applications.
