Japan to issue cybersecurity guidelines for advanced AI tools

Japan will draw up guidelines to bolster cybersecurity in response to the emergence of powerful AI tools like Anthropic's Claude Mythos model, with the government planning to urge software providers to use these programs to check for vulnerabilities. At the same time, Anthropic is set to discuss cyber vulnerabilities in the global financial system exposed by its latest Mythos model with members of the Financial Stability Board, according to a Financial Times report citing people familiar with the plan, though Reuters could not immediately verify the information. The Claude Mythos model is capable of discovering vulnerabilities in major operating systems, a capability that has prompted Anthropic to restrict access to the tool due to the cybersecurity risks it poses. Japan's response reflects growing concern among governments and financial regulators about the dual-use nature of advanced AI systems that can enhance cybersecurity detection while simultaneously presenting potential threats if misused. These developments come as authorities worldwide grapple with how to regulate and manage powerful AI tools that can be leveraged both for defensive cybersecurity purposes and potentially for malicious activities. The planned discussions with the Financial Stability Board highlight the broader systemic risks that such AI capabilities could pose to global financial infrastructure if mishandled.