IMF Warns AI Cyberattacks Pose Systemic Global Financial Risk

The International Monetary Fund warned in May 2026 that AI-powered cyberattacks pose a systemic risk to global financial stability. In a report and subsequent blog post, the organization stated that extreme cyber-incident losses could trigger funding strains, raise solvency concerns, and disrupt broader markets. The IMF highlighted that the financial sector's reliance on a small number of cloud providers increases the risk of contagion, potentially leading to a worldwide financial crisis by disrupting payments and liquidity.

These warnings focused on the capabilities of Anthropic's Mythos model, which the IMF described as being able to identify and exploit vulnerabilities in major operating systems and browsers at machine speed. To mitigate these risks, Anthropic established Project Glasswing, limiting the model's distribution to 40 critical companies, including Microsoft, Apple, and Nvidia, rather than a general release.

In response to the escalating threat, the United States government shifted its policy to evaluate new AI models from technology companies before they are released. White House economic adviser Kevin Hassett announced a joint all-government and private sector effort to test the Mythos model to prevent harm to U.S. businesses. The IMF emphasized that emerging and developing economies are disproportionately exposed due to resource constraints and called for urgent international coordination, cyber stress testing, and board-level oversight to preserve global stability.