U.S. Regulators Pause Cybersecurity Bank Exams Due to Mythos

Source Domain: www.govtech.com

(TNS) — U.S. regulators are pausing some cyber-related examinations of the largest banks, giving the lenders more breathing room as the firms dig into the risks exposed by Anthropic PBC’s new Mythos AI model.

The Federal Reserve and Office of the Comptroller of the Currency want to give banks time to bolster their systems against cyber threats exposed by the latest AI models as both regulators and banks test the new technology, said people familiar with the matter, who asked not to be identified citing private information.

A representative for the OCC declined to comment, while the Fed had no immediate comment.

Anthropic said last month that it would limit access to Mythos, warning that it could potentially be used to power cyber attacks. As part of that, the tech company created an initiative called Project Glasswing with a select few companies, including Apple Inc. and JPMorgan Chase & Co., that would be allowed to use the technology to test their own cyber defenses first.

In April, Treasury Secretary Scott Bessent and then-Fed Chair Jerome Powell summoned Wall Street leaders to warn them of the cyber risks raised by AI models such as Mythos. Since then, banks and officials have been hashing out ways to methodically tackle the vulnerabilities that Mythos has exposed, according to people familiar with the matter.

Inside the big U.S. banks, some executives were immediately alarmed by Mythos’s ability to move quickly through code and identify hacking weaknesses. But after several weeks of testing, some of the initial panic has given way to a long to-do list.

Delaying some cyber exam schedules will give banks more time to fully understand the power of the new technology and will help regulators with ongoing efforts to stress-test defenses for it. The OCC, for example, is undergoing its own trial run with Mythos, one of the people said. To be sure, examiners are still engaged…

Source