Artificial Intelligence

Anthropic to Give UK Banks Access to Mythos AI Model

Is Claude Mythos an unprecedented security crisis or a responsible defense effort that puts defenders ahead?

Published 1 hour ago

Anthropic to Give UK Banks Access to Mythos AI Model

story

Above: Pip White of Anthropic during a Bloomberg Television interview in London, U.K., on April 16. Image credit: Chris Ratcliffe/Bloomberg/Getty Images

The Spin

Establishment-critical narrative

Claude Mythos is a genuine turning point in global security, the kind that doesn't wait for governments to catch up. Anthropic's own testing found thousands of critical vulnerabilities across every major operating system and browser, some hidden for nearly three decades, meaning the model could be used for large-scale cyber exploitation. U.K. and U.S. regulators are already in emergency talks with major banks, and treasury leaders who aren't treating this as a top strategic priority right now are dangerously behind.

Danny Kruger on X The Global Treasurer The New York Times

Pro-establishment narrative

Anthropic expanding Mythos access to U.K. banks, following a limited release to U.S. companies, reflects a controlled rollout rather than broad deployment. The company is withholding full public release, committing $100M in model credits to patch vulnerabilities, and partnering with leading tech and financial institutions via Project Glasswing to harden critical infrastructure. The approach aims to ensure defensive safeguards improve alongside capability, with early adopters helping strengthen system security before wider release.

Anthropic on X Anthropic