Finance leaders warn over Mythos as UK banks prepare to use powerful Anthropic AI tool | AI (artificial intelligence)
British banks will next week be given access to a powerful artificial intelligence tool deemed too dangerous to be made public, as a number of senior financial officials have warned of its impact.
Anthropic, which has so far limited the launch of the new model to a small group of US businesses including Amazon, Apple and Microsoft, said it would expand it to financial institutions in the UK.
“This will happen in the very near future, next week,” Pip White, Anthropic’s head of operations for the UK, Ireland and Northern Europe, said in an interview with Bloomberg TV. “As you would expect, the attention I received from UK CEOs last week was quite significant.”
Anthropic, the company behind the Claude family of AI tools, said its latest model, Mythos, poses an unprecedented risk due to its ability to reveal flaws in IT systems.
“AI models have reached a coding capacity that can outpace all but the most skilled humans at finding and exploiting software vulnerabilities,” Anthropic said in a blog post earlier this month. “The consequences for economies, public safety and national security could be serious.”
Finance ministers, administrators and regulators discussed potential threats when they gathered in Washington this week for the spring meetings of the IMF and World Bank, while also addressing concerns about the global consequences of the U.S.-Israeli war with Iran.
Canadian finance minister François-Philippe Champagne told the BBC: “It is undoubtedly serious enough to attract the attention of all finance ministers… The difference with the Strait of Hormuz is that we know where it is and how big it is.
“The problem we face with Anthropic is that it’s an unknown unknown. It requires a lot of attention so that we can put safeguards in place, and we have processes in place to make sure that we ensure the resilience of our financial system.”
Bank of England Governor Andrew Bailey, who also chairs the Financial Stability Board of global regulators, said: “This is a very serious challenge for all of us. It reminds us how fast the world of AI moves.”
But as governments try to reap the economic benefits of artificial intelligence, regulators have to consider whether and how difficult it would be to restrict the technology, he said. “What is the best moment to frame the rules of the road?” Bailey asked. “If you go too early you run the risk of a) missing the target and b) disrupting the evolution, and if you go too late things can get out of control.”
European Central Bank president Christine Lagarde said: “The development we’ve seen with Anthropic and Mythos is a good example of a responsible company suddenly thinking: ‘Oh, this could be really good’ – but if it falls into the wrong hands it could be really bad.”
“Everyone wants to have a framework to operate,” Lagarde told Bloomberg TV. But he added: “I don’t think there’s a governance framework that really cares about these things. We need to work on that.”
US treasury secretary Scott Bessent summoned US bank bosses to Washington last week to discuss the Mythos model. This meeting focused on systemically important banks where regulators believe a major disruption of their operations or potential collapse would put financial stability at risk.
UK regulators will discuss the risks of Mythos with bank bosses and government officials in the coming weeks.
Dan Katz, IMF deputy managing director and former chief of staff to Bessent, said: “The evolution of digital technology poses enormous risks to cybersecurity… this will be absolutely crucial on the international agenda over the next few months.”
