UK banks still lack access to Mythos AI model, BoE's Bailey says

British banks have yet to gain access to Anthropic's Mythos artificial intelligence model, intended to bolster their defenses against cyber threats, according to Bank of England Governor Andrew Bailey. The AI model first raised concerns six weeks ago, but access has been stalled. Bailey indicated that while Anthropic is willing to share the models for trial, a "political hold-up" involving the U.S. administration is preventing progress. He expressed his inability to fully explain the differing processes for various companies but stressed the importance of access given the perceived risks. Anthropic has previously faced scrutiny from the U.S. administration regarding the potential military applications of its AI tools. Some cybersecurity experts, however, have downplayed fears of widespread hacking facilitated by the model. The situation unfolds as President Donald Trump recently postponed a broader executive order on AI development. Bailey, who also leads the Financial Stability Board, underscored the necessity of a unified global strategy for addressing cyber risks. He warned that isolated national approaches would be ineffective due to the significant potential for cross-border "spillovers" from cyber threats, stating that even banks believing they have secured their systems would be vulnerable due to the interconnected nature of global finance.

“

It hasn't happened yet and I think this has been somewhat caught up in the process with the U.S. administration.