Japan eyes using cutting-edge AI under preemptive cyber defense policy

Japan is contemplating the use of advanced artificial intelligence (AI) systems within its preemptive cyber defense policies, in response to the potential threat posed by sophisticated models like Anthropic's Claude Mythos, a senior government official has disclosed. Concerns over cyberattacks leveraging powerful AI models with the ability to identify security weaknesses have prompted Japan to embrace such technology, National Cyber Director Yoichi Iida stated in a recent interview with Kyodo News.

Iida warned that the time between identifying a vulnerability and exploiting it could drastically reduce, possibly to a fraction of a hundredth or thousandth of its previous duration. The Claude Mythos AI model, known for its exceptional capability in detecting security vulnerabilities, initially restricted access to IT companies like Google and select financial institutions due to potential misuse for cyberattacks.

Prime Minister Sanae Takaichi's Cabinet, in March, authorized the government to initiate active cyber defense measures from October 1 to mitigate the risk of serious cyberattacks that could disrupt critical infrastructure, economic activities, and people's livelihoods. Iida emphasized the need for Japan to consider advancing automation in its cyber defense operations while maintaining human oversight in decision-making.

He cautioned against relying on a single AI model, citing the U.S. government's export control directive that suspended access to Mythos-class models in June. This move impacted security vulnerability scanning operations for crucial government systems. Iida warned that if such cutting-edge AI models were manipulated in nation-sponsored cyberattacks, the attacks could become far more intricate, with increased intensity.