US Regulators Pause Bank Cyber Exams to Address Mythos Vulnerabilities
May 19, 2026 – 9:54 pm
Image by: Shutterstock
TL;DR
US banking regulators are temporarily pausing some cyber security examinations of the largest banks to give them time to address vulnerabilities exposed by Anthropic’s AI model, Mythos. This comes after weeks of testing revealed thousands of zero-day flaws across major operating systems and browsers, raising concerns about potential sophisticated cyberattacks.
Background
Weeks after Anthropic’s Mythos AI model emerged, causing a stir in the financial sector, America’s top banking regulators have made a strategic decision. The Federal Reserve and the Office of the Comptroller of the Currency (OCC) are suspending certain cyber-related examinations of major US banks.
Official Statement: This pause is not a retreat from oversight but a deliberate move to allow institutions to focus on patching vulnerabilities while they conduct their own Mythos trials.
Initial Impact
Initially, Wall Street expressed alarm when Anthropic restricted access to its new frontier model due to the potential threat it posed. Under Project Glasswing, banks discovered Mythos’ remarkable ability to identify hacking weaknesses at an unprecedented speed—a capability that prompted a critical meeting between senior leaders from Goldman Sachs, Morgan Stanley, Citigroup, Bank of America, and Wells Fargo with Treasury Secretary Scott Bessent and then-Fed Chair Jerome Powell.
Current Status
After extensive testing, the initial panic has subsided, giving way to a detailed to-do list. Major banks with access to Mythos, including JPMorgan Chase, Morgan Stanley, and Goldman Sachs, have assembled dedicated teams to assess and fix the vulnerabilities. Many are also collaborating with federal intelligence agencies to better understand the broader threat landscape.
JPMorgan CEO Jamie Dimon’s Quote: "It’s serious work. We have, I think, hundreds of people doing it full time now."
Goldman Sachs CEO David Solomon’s Comment: "We are working closely with Anthropic and all of our security vendors."
Meanwhile, the OCC is conducting its own trial run with the model to gain firsthand understanding before resuming bank examinations.