Japan plans to deploy advanced AI for proactive cyber defence amid rising threats

Japan is considering integrating state-of-the-art AI systems into its pre-emptive cyber-defence framework, according to one outlet Cyber Director Yoichi Iida. He highlighted that models such as Anthropic’s Claude Mythos are capable of rapidly identifying software vulnerabilities, potentially shrinking the window between discovery and exploitation to a fraction of its former length. The discussion follows Prime Minister Sanae Takaichi’s cabinet resolution to enable “active cyber defence” measures starting Oct. 1, aimed at averting serious attacks on essential services and the economy.

Iida said the government must advance automation in cyber-operations but retain human oversight for final actions. He also cautioned against depending on a single AI model, noting the U.S. government’s recent export-control directive that suspended access to Mythos-class models, which disrupted vulnerability-scanning for key systems. The official warned that nation-state actors could exploit advanced AI to launch far more sophisticated cyber offensives.