UK banks locked out of Mythos a ‘wake-up call’, adviser says
TL;DR:
- The Treasury’s banking AI champion says UK lenders’ inability to access Anthropic’s Mythos model underlines the urgency of building British AI capability.
- Only a small number of UK banks — mainly the local operations of US lenders — were granted access to the cybersecurity-focused model.
- Recommendations published alongside the government’s AI Adoption Plan urge regulators to review AI chatbots giving financial advice.
British banks’ exclusion from Anthropic’s Mythos model has become the sharpest available argument for building sovereign AI capability, according to the government adviser appointed to lead the sector’s thinking on the technology. Harriet Rees, Starling Bank’s chief information officer and the finance ministry’s “AI champion” since January, told Reuters that Britain must develop its own infrastructure, models and skills rather than rely wholly on US technology.
No timeline, and no time
The commercial stakes are specific. Banks want Mythos because it is regarded as the most advanced model for identifying cybersecurity vulnerabilities, letting them patch weaknesses faster. Anthropic released it in April to a select group including JPMorgan, and in Britain only a very small number of banks — mainly the UK arms of US lenders — were granted access. The chief executive of one of Britain’s largest banks said last week they still have no timeline for when access might arrive.
“Now is the time for us to think very strategically about what we need to do to protect our leading position moving forward,” Rees said. “Time really is of the essence; we don’t have two years here.”
That urgency now has an official register. The exclusion follows US restrictions on who can access the model, and lands as the Bank of England presses for international coordination on frontier AI and the chancellor stakes her growth argument on AI sovereignty. It also gives concrete shape to a risk UK regulators have described in the abstract: the Bank has already flagged AI as a financial stability concern, and dependence on a handful of foreign suppliers is precisely the concentration it worries about.
Rees and fellow appointee Rohit Dhawan, Lloyds Banking Group’s head of AI, published recommendations alongside the government’s financial services AI Adoption Plan, urging regulators to review the growing use of AI chatbots to deliver financial advice. Dhawan suggested the UK could designate the AI companies banks already depend on as “critical” providers, bringing them within the remit of financial regulators — joining four US cloud providers named last week. Rees added that Britain should widen its options by building relationships with AI companies outside the US, including in China and France.
Looking forward
The government said it will work with regulators and industry on next steps. For UK firms watching from outside financial services, the episode is a useful test of what sovereignty means in practice: not whether Britain can build a frontier lab, but whether it can guarantee its own banks access to the tools defending them.