UK banks' lack of Mythos access a wake-up call, government AI adviser says

• Mythos is considered the most advanced AI model for identifying cybersecurity vulnerabilities, but only UK operations of U.S. lenders have gained access so far
• Government adviser Harriet Rees warns that 'time is of the essence' and Britain doesn't have 'two years' to address AI sovereignty issues
• New recommendations include bringing AI firms under financial regulatory oversight as 'critical' providers and building relationships with non-U.S. AI companies in China and France

Summary

UK banks' inability to access Anthropic's advanced AI model Mythos has triggered concerns about Britain's reliance on U.S. technology providers, according to government banking adviser Harriet Rees, CIO at Starling Bank. Anthropic released Mythos in April to select organizations including U.S. lender JPMorgan, but only UK operations of American banks gained access in Britain.

Key Issue: Mythos is considered the most advanced AI model for identifying cybersecurity vulnerabilities, making it critical for banks to strengthen defenses and patch weaknesses quickly. Major British lenders still lack a timeline for access, while Anthropic has begun limited international rollout beyond the U.S.

Strategic Response: Rees, appointed by the finance ministry in January alongside Lloyds Banking Group's AI head Rohit Dhawan, emphasized the urgency: "Time really is of the essence, we don't have two years here." Their recommendations include:

• Building domestic AI infrastructure, models, and skills
• Developing relationships with non-U.S. AI firms, including those in China and France
• Designating critical AI providers as "critical" to the finance sector under regulatory oversight
• Reviewing AI chatbot use for financial advice delivery

Government Action: The proposals were published Tuesday as part of the government's financial services AI adoption plan, ahead of Finance Minister Rachel Reeves' annual Mansion House speech. Reeves stated she has "set out a serious plan for AI sovereignty, backing British companies to compete and win."

Regulators have warned that concentration among third-party technology providers poses financial stability and operational risks to the sector.