Anthropic Plans to Bring Mythos to UK Banks Within the Next Week

Anthropic PBC is planning to release its closely watched Mythos artificial intelligence model to UK financial institutions in the coming week.

The tech company is gradually expanding "Project Glasswing," a program to give select organizations early access to the AI, after discovering that the model is a powerful tool for spotting and potentially exploiting cybersecurity vulnerabilities.

"That is in the very near term, in the next week," said Pip White, Anthropic's head for the UK, Ireland and northern Europe, in an interview on Bloomberg Television. "As you would expect, the engagement I have had from UK CEOs in the last week has been significant."

Anthropic said earlier this month that it would limit the release of the AI model to give cybersecurity professionals a chance to test it against their own defences. According to Anthropic, Mythos testing has already found thousands of "zero-day" vulnerabilities during testing, including in every major operating system and every major web browser.

The initial group in the Glasswing program included large tech companies: Amazon.com Inc., Apple Inc., Microsoft Corp., Cisco Systems Inc. Soon after Glasswing was announced, US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell summoned Wall Street leaders for an urgent discussion about Anthropic's Mythos and similar AI models.

The lack of access and limited knowledge about the capabilities of the model has sparked concerns among banks and government agencies. Bank of England Governor Andrew Bailey said on Tuesday that global regulators need to rapidly evaluate the threat posed by Anthropic's model.

"We are putting our own safeguards and our own limitations around this product because we know how powerful it can be," White said.

The UK's AI Security Institute said on Wednesday that it had access. The government-backed institution in charge of evaluating AI risks found that the model was "a step up over previous frontier models" in simulating multistep cyber attacks.

White also said Anthropic plans to open a London office in the first quarter of next year. The company, which has around 200 employees in London, said the new office will give them the capacity for 800.