Government to urge software developers to use powerful AI to check for vulnabilities
Anthropic has restricted access to Claude Mythos, a powerful AI tool that can discover vulnerabilities in major operating systems, over the cybersecurity risks it poses. (Photo by Suzu Takahashi)
SATOSHI TEZUKA
May 18, 2026 02:31 JST
TOKYO -- Japan will draw up guidelines on how to bolster cybersecurity with Anthropic's Claude Mythos model, including such steps as requesting software providers to check for vulnerabilities using the powerful AI tool.