British financial authorities are conducting emergency discussions with cybersecurity officials and banking leaders to evaluate potential threats from a new artificial intelligence system developed by Anthropic, according to a Financial Times report published Sunday.
The Bank of England, Financial Conduct Authority, and HM Treasury have initiated conversations with the National Cyber Security Centre to study possible weaknesses in essential computer systems that Anthropic’s newest AI technology has exposed, the publication stated.
Banking executives, insurance company leaders, and stock exchange officials from major British institutions are scheduled to receive briefings about cybersecurity dangers associated with the AI system, known as Claude Mythos Preview, during regulatory meetings planned for the coming two weeks, sources familiar with the discussions told the newspaper.
Reuters was unable to independently confirm the Financial Times report. Anthropic has not responded to requests for comment, while the Bank of England refused to provide statements and other agencies were unavailable for immediate response.
These developments come after U.S. Treasury Secretary Scott Bessent convened similar discussions with prominent Wall Street financial institutions regarding the AI model’s cybersecurity implications, Reuters reported Friday based on source information.
The artificial intelligence company has explained that the system is being implemented through “Project Glasswing,” a restricted program allowing selected organizations to utilize the unreleased Claude Mythos Preview model specifically for protective cybersecurity applications.
According to a company blog post published earlier this month, the AI system has successfully detected “thousands” of significant security vulnerabilities in operating systems, internet browsers, and other commonly utilized software programs.
