British banks will gain access within the next week to Anthropic's powerful new AI model, Mythos, which the company has deemed too dangerous for public release. The tool, currently limited to a handful of US firms including Amazon, Apple, and Microsoft, is being expanded to UK financial institutions, according to Pip White, Anthropic’s head of UK, Ireland and northern Europe operations.
Anthropic has warned that Mythos poses unprecedented risks due to its ability to expose flaws in IT systems. In a blogpost, the company stated that AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities, with potential severe fallout for economies, public safety, and national security.
Finance ministers and regulators gathered in Washington for the IMF and World Bank spring meetings discussed the threats. Canadian finance minister François-Philippe Champagne described the risk as an 'unknown unknown' requiring safeguards. Bank of England governor Andrew Bailey called it a very serious challenge, noting the need to balance regulation with fostering AI's economic benefits.
European Central Bank president Christine Lagarde highlighted the need for a governance framework, while US treasury secretary Scott Bessent held a meeting with US bank bosses on the issue. UK regulators are expected to raise the matter with bank bosses and government officials in the coming weeks.
