UK Regulators Call Emergency Meetings Over Anthropic's Hacking Model

April 13, 2026

UK Regulators Call Emergency Meetings Over Anthropic's Hacking Model

Published: April 13, 2026 at 12:31 AM

Updated: April 13, 2026 at 12:31 AM

100-word summary

Anthropic's latest model is so good at finding security flaws that UK financial regulators are treating it like a national security event. Claude Mythos Preview has already uncovered thousands of vulnerabilities in major operating systems and browsers. Now the Bank of England, Financial Conduct Authority, and National Cyber Security Centre are rushing to brief banks and insurers on the risks before wider release. The model is currently locked down under Project Glasswing, which lets select organizations use it only for defense. US Treasury Secretary Scott Bessent held similar emergency talks with Wall Street banks. The irony: a tool built to protect systems is dangerous enough to require its own controlled...

What happened

Anthropic's latest model is so good at finding security flaws that UK financial regulators are treating it like a national security event. Claude Mythos Preview has already uncovered thousands of vulnerabilities in major operating systems and browsers. Now the Bank of England, Financial Conduct Authority, and National Cyber Security Centre are rushing to brief banks and insurers on the risks before wider release. The model is currently locked down under Project Glasswing, which lets select organizations use it only for defense. US Treasury Secretary Scott Bessent held similar emergency talks with Wall Street banks.

Why it matters

The irony: a tool built to protect systems is dangerous enough to require its own controlled deployment program.

Sources