TOKYO -- Japan will draw up guidelines on how to bolster cybersecurity in response to the emerge of such powerful AI tool as Anthropic's Claude Mythos model, urging software providers to use those programs to check for vulnerabilities.