Private sector urged to ‘fight AI with AI’ to defend infrastructure

Source Domain: www.asahi.com

Japan is urging systems developers of critical infrastructure to utilize advanced AI resources, such as the recently developed Claude Mythos, to swiftly identify and mitigate vulnerabilities in potential cyberattacks.
The initiative, outlined at a pivotal inter-ministerial meeting on May 18, aims to combat the escalating threats posed by “frontier AI models.”
Minister in charge of cybersecurity, Hisashi Matsumoto, emphasized the urgency of escalating government efforts to counter the increasing speed of cyberattacks.
The Japanese government is seeking partnerships with the U.S. to grant access to Mythos, a model developed by U.S. startup Anthropic, which has been withheld from the public due to potential misuse.
Under the “Project YATA-Shield,” the government plans to raise cybersecurity awareness among key infrastructure operators, share cyberattack information across industries, and expand public-private partnerships.
The government is also strengthening collaborations with foreign agencies and AI developers, and will work closely with the Japan AI Safety Institute to update guidelines as necessary.
The meeting included officials from various relevant ministries and agencies, focusing on key sectors such as finance, transportation, and electric power.

You may have missed