IMF chief warns advanced AI models like Mythos “can be used to destroy the financial system”

The head of the International Monetary Fund warned on Wednesday that advanced AI models like Anthropic’s Mythos “can be used to destroy the financial system” if they fall into the wrong hands. IMF Managing Director Kristalina Georgieva made the comments in Brussels while presenting the Fund’s annual economic assessment of the eurozone.

“What we recognize is that Mythos is just the beginning, there will be more like it,” Georgieva told journalists. She said the same capabilities that allow frontier models to identify cybersecurity vulnerabilities “with speed and scale that was unthinkable until now” can also be weaponised. “In the wrong hands, that same capacity can be used to destroy the financial system.”

Georgieva called for countries to have “all the elements of cybersecurity in place” and urged more cooperation between rich and poor nations, and between the public and private sectors. She noted there is no world cybersecurity organisation and that “in the current geopolitical environment it’s hard to imagine that we can have one.”

Advanced economies “have to figure out a way to help the developing world” defend itself, she said, because the global financial system is too interconnected to tolerate weak links. “If there is a big weakness, it will be utilized.” She also told countries to factor AI-related cybersecurity into their budgets. “Recognize that patching costs money. Make sure that you have the fiscal space.”

The warning comes as central banks and regulators scramble to respond to the Mythos risk. The Fed and Treasury have convened bank CEOs. The ECB is running cybersecurity exercises. Euro-area finance ministers have demanded access to Mythos after discovering no EU institution can see what the model finds. Japan, Australia, and the UK have all sought briefings.

Georgieva also flagged the risk of the AI investment boom turning into a bust. “We also could see the AI enthusiasm, the AI boom more at risk of turning into an AI bust,” she said. “We don’t see it as high probability, but it’s also not zero. So it falls in the category of low probability, very high impact risks.”

That warning lands the same day SpaceX debuted at a $2 trillion valuation, Amazon borrowed $17.5 billion for AI infrastructure, and Alphabet’s $85 billion equity offering closed. The total AI-related capital commitment from Big Tech in 2026 exceeds $700 billion. Georgieva is saying, diplomatically, that the financial system is not prepared for a world where that spending does not pay off.

The eurozone’s own outlook has weakened. The IMF projects growth of 0.9% in 2026, down 0.5 percentage points from pre-war estimates. Inflation is rising to a projected 2.8%. Georgieva said 80% of government measures to help households cope with energy costs do not meet the IMF’s standard of being “temporary and targeted.” Against that backdrop, asking countries to find fiscal space for AI cybersecurity is a hard sell, but one the IMF is making anyway.