UK Regulators Assess Risks of Anthropic's New AI Model

UK regulators -- including the Bank of England, FCA, Treasury and NCSC -- are urgently convening with major banks to assess cyber and systemic risks posed by Anthropic's powerful new AI model, Claude Mythos. This follows high-level concern over its unprecedented capabilities.

UK financial regulators rush to assess risks of Anthropic's latest AI model, FT reports

Urgent Regulatory Response to Anthropic's AI Model

April 12 (Reuters) - UK financial regulators are holding urgent talks with the government's cyber security agency and major banks to assess risks posed by the latest artificial intelligence model from Anthropic, the Financial Times reported on Sunday.

Collaboration Among UK Financial Authorities

Officials at the Bank of England, the Financial Conduct Authority and HM Treasury are in talks with the National Cyber Security Centre to examine potential vulnerabilities in critical IT systems highlighted by Anthropic's latest AI model, the newspaper said.

Engagement with Major Financial Institutions

Representatives from major British banks, insurers and exchanges are expected to be briefed on the cyber security risks posed by the model, Claude Mythos Preview, at a meeting with regulators in the next fortnight, it said, citing two people briefed on the talks.

Verification and Official Responses

Reuters could not immediately verify the report. Anthropic did not respond to Reuters' request for comment while the BoE declined to comment and the Treasury, NCSC and FCA were not immediately available for comment.

International Context and Project Details

The move follows a meeting called by U.S. Treasury Secretary Scott Bessent with major Wall Street banks on the model's cyber risk potential, Reuters nL4N40T024 reported on Friday, citing two sources.

Project Glasswing and Model Deployment

The AI startup has said the model is being deployed as part of "Project Glasswing nL4N40Q0LK", a controlled initiative under which select organizations are permitted to use the unreleased Claude Mythos Preview model for defensive cyber security purposes.

Early Findings and Vulnerability Identification

In a blog post earlier this month, the startup said the model had already identified "thousands" of major vulnerabilities across operating systems, web browsers and other widely used software.

(Reporting by Mihika Sharma in BengaluruEditing by Bernadette Baum and Christina Fincher)