UK banks' restricted access to Anthropic's advanced AI model, Mythos, underscores the urgent need for Britain to bolster its domestic artificial intelligence capabilities, according to Harriet Rees, an AI champion appointed by the finance ministry.
Rees, who is also the chief information officer at Starling Bank, stated that the current situation emphasizes the necessity for a strategic approach to AI development, including building essential infrastructure and models to avoid over-reliance on U.S. technology. She stressed that "time really is of the essence" for the UK to act.
Mythos, released by U.S.-based Anthropic in April, is recognized for its ability to identify cybersecurity vulnerabilities. However, access has been limited, with only a small number of UK banks, primarily the domestic operations of U.S. lenders, reportedly gaining entry. Other major British banks are awaiting a timeline for access.
Anthropic has indicated it has begun rolling out Mythos 5 to organizations outside the United States and is working with the U.S. government to broaden access. Bank of England Governor Andrew Bailey has also called for international cooperation on managing the risks associated with advanced AI models.
In response, Rees and fellow appointee Rohit Dhawan have developed a set of recommendations focused on AI policy and regulation to encourage greater adoption of the technology within the financial services sector. These proposals were released as part of the government's AI adoption plan, with Finance Minister Rachel Reeves affirming a commitment to "AI sovereignty" and supporting British companies in the global AI landscape.
