Global banking crisis ahead? Anthropic's Mythos AI sparks fresh concerns

Claude Mythos Preview AI model can identify vulnerabilities in lenders' cyber defences

Senior international financial officials, including finance ministers and central bankers, have raised fresh concerns regarding the "unprecedented" impacts of the latest AI models on the world banking system.

At the heart of the concern lies Anthropic's newly-released Anthropic Claude Mythos Preview, capable of identifying vulnerabilities in every major operating system and web browsers.

According to top financial leaders, Claude Mythos can severely disrupt the global financial system by exposing weaknesses in lenders' cyber defence.

Andrew Bailey, governor of the Bank of England, who chairs the Financial Stability Board of global regulators, said, "It is a very serious challenge for all of us. It reminds us how fast the AI world moves."

Bailey also issued a warning regarding the development of powerful AI models, making it easier to detect the existing vulnerabilities in IT systems. If these models have fallen to rogue actors, they can easily be exploited.

He also urged the global regulators to evaluate the potential cyber security threats posed by Mythos AI model.

Similar disturbing concerns were also raised by the top officials belonging to the IMF and World Bank. According to Dan Katz, deputy head of the IMF and former chief of staff for US Treasury secretary Scott Bessent, with the rapid evolution of digital technologies, specifically AI models, cyber security has become the need of the hour, discussing it on the international agenda.

European Central Bank president Christine Lagarde said, "The development we've seen with Anthropic and Mythos is a good example of a responsible company that is suddenly thinking, 'ah, that could be really good' -- but if it falls in the wrong hands, it could be really bad."

Canadian Finance Minister François-Philippe Champagne also presented the more bleak picture of risks posed by Mythos AI models. According to him, the recent crises driven by the Iran conflict are still known to us, but when it comes to Anthropic's Claude Mythos, we are facing an unknown threat.

"Certainly it is serious enough to warrant the attention of all the finance ministers. It requires a lot of attention so that we have safeguards, and we have processes in place to make sure that we ensure the resiliency of our financial system," Champagne added.

James Wise, a partner at Balderton Capital, is chair of the Sovereign AI unit, called Mythos "the first of what will be many more powerful models."

Anthropic issued a stark warning that these capabilities will soon spread, likely reaching groups that lack a commitment to safe deployment.

"The fallout -- for economies, public safety and national security -- could be severe," the company cautioned.

Given Mythos unprecedented cyber capabilities, Anthropic decided not to release this model to the public. Under project glasswing, Mythos will be available to tech giants, Google, Microsoft, and 40 other organizations.