Banks in the United Kingdom are still unable to access Anthropic’s Mythos model to test their systems against cyber threats, Bank of England Governor Andrew Bailey said Friday. Bailey’s comment comes six weeks after the model first drew regulatory concern. Bailey said Anthropic was willing to share Mythos on a trial basis, but the rollout had stalled.

“It hasn’t happened yet, and I think this has been somewhat caught up in the process with the U.S. administration,” Bailey said on the sidelines of the central banking conference in Reykjavik. As Cryptopolitan reported in April, the Bank of England, the Financial Conduct Authority, HM Treasury, and the National Cyber Security Centre convened to assess the risks Mythos posed to British financial institutions. At the time, Anthropic’s head of UK, Ireland, and Northern Europe, Pip White, said that UK banks would receive access to Mythos within the week.

Bailey comments on issues with Mythos rollout

Bailey named Mythos explicitly in an April 15 speech at Columbia University, describing it as a major cybersecurity concern and saying cyber had climbed regulators’ risk rankings faster than any other category in recent years. Mythos is currently limited to a select few companies through Anthropic’s Project Glasswing for cybersecurity applications. Early access was given to Goldman Sachs and a few other US companies.

Cryptocurrency firms and UK banks have been left out of the first release. Bailey, who also chairs the International Financial Stability Board, said cyber threats cannot be contained within national boundaries. “Spillovers from this sort of cyber risk are so big that we can’t just have a single sort of national approach,” he said. Banks are deeply interconnected across borders, he added, meaning that one country securing its own institutions would not be enough if others remain exposed.

Anthropic has said the model can find and exploit software vulnerabilities better than all but the most skilled human experts. When Anthropic released Mythos to select customers, the company said it had already found thousands of high-severity vulnerabilities across open- and closed-source software, more than 99% of which remained unpatched. In the hands of defenders, Mythos could let banks find and fix flaws before attackers reach them.

Outside that circle, the same capability becomes a threat. The banks still waiting for access are exposed to a tool that their potential attackers may eventually obtain. Bailey’s comments come while Anthropic is also at odds with the Trump administration over military access to its AI tools. The dispute is about where to draw the line on how the U.S. military can use the company’s technology.

President Donald Trump recently postponed signing an executive order on artificial intelligence that would have created a voluntary process in which developers could seek input from the federal government before making their advanced models public. That postponement adds another layer of uncertainty for companies and regulators seeking clearer rules around frontier AI systems.

As for British banks, despite being flagged as exposed by their own central bank, they remain on the outside of a tool that the Bank of England considers important enough to raise at an international conference. Anthropic has not publicly detailed the specific hold-up. The company said in April it was prepared to begin offering Mythos to British banks, with White citing “significant” engagement with UK bank chief executives since the model’s release.

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.