Japan has asked ANTHROPIC and OpenAI to grant broader access to advanced AI
models that can detect cybersecurity vulnerabilities and will encourage
critical-infrastructure operators to adopt the tools. The measure will be part
of a cybersecurity package slated as early as next week to deepen cooperation
with US AI developers. Prime Minister TAKAICHI on Monday directed cabinet
members to push the work and to cooperate with foreign agencies to identify
attackers and remediate system flaws amid a recent wave of data breaches. On
Oct. 13, Japan’s cybersecurity chief Shunji Furukawa met ANTHROPIC co-founder
Jack Clark; ANTHROPIC agreed to work with a government AI-security research body
to analyze methods for AI-enabled attacks.
