Singapore Calls for AI to Fight AI as Cyber Threats Grow More Sophisticated
Singapore is calling for a multi-layered defence against increasingly capable artificial intelligence, including using AI itself to detect cyber weaknesses and counter AI-driven attacks.
Image: Tech360TV (AI-generated illustration)
Minister for Digital Development and Information Josephine Teo said no single safeguard would be enough as AI systems become more autonomous. Singapore’s approach combines stronger cyber defences, clear limits on what AI can do, human oversight and continuous testing of advanced models.
Teo said organisations should use available AI tools defensively, such as to identify vulnerabilities before attackers can exploit them. The Cyber Security Agency of Singapore has also urged organisations to patch systems quickly and monitor networks for unusual behaviour.
The concern is growing as AI agents are given more freedom to carry out tasks with limited human supervision. Such systems can misunderstand instructions, be manipulated by malicious content or take unintended actions, creating new risks for businesses and public-sector systems.
Singapore is also pushing for stronger international cooperation on AI safety. The country has joined other governments in calling for common safeguards for frontier AI models and is working with developers, researchers and independent testers to compare findings on emerging risks.
Locally, IMDA’s Model AI Governance Framework for Agentic AI recommends safeguards such as monitoring agent actions and requiring human approval for higher-risk tasks. IMDA, CSA and GovTech are also working with industry partners to test AI agents in isolated environments.
