Anthropic PBC is set to release its Mythos artificial intelligence model to UK financial institutions in the coming weeks under its “Project Glasswing” programme, which provides limited early access to selected organisations, as confirmed by Pip White, the company’s UK, Ireland and Northern Europe chief, in a recent interview with Bloomberg.

The model has reportedly uncovered thousands of zero-day vulnerabilities across major operating systems and browsers during testing, raising concerns over its cybersecurity implications.

UK and international financial and regulatory attention has intensified ahead of the rollout. White said interest from UK chief executives has been significant and that safeguards are being applied due to the system’s capabilities. Early access participants have so far included Amazon, Apple, Microsoft and Cisco.

The development has prompted discussions among policymakers, with US and UK officials and regulators examining potential risks, including the Bank of England.

The UK AI Security Institute said the model represents a step up from previous frontier systems in simulating cyber attacks.

Anthropic also plans to expand its London office capacity substantially next year.