Harriet Rees, chief information officer at Starling Bank and the finance ministry's designated AI champion, said the difficulty UK banks have faced in obtaining access to Anthropic's Mythos model demonstrates the urgency for Britain to develop its own AI capabilities. Rees argued that the country needs to create infrastructure and build models while training the necessary skills so that financial institutions are not fully dependent on U.S.-based technology providers.
"Now is the time for us to think very strategically about what we need to do to protect our leading position moving forward. Time really is of the essence, we don’t have two years here," Rees said, stressing the limited window to act.
Anthropic released Mythos in April to a select group of organisations, among them the U.S. bank JPMorgan. The model is prized by banks for its capabilities in detecting cybersecurity weaknesses, enabling faster patching and enhancing defensive measures. In the UK, only a very small number of banks - primarily the local operations of U.S. lenders - were granted access, according to Rees.
Senior executives at major British banks remain uncertain when they might obtain access. The chief executive of one of the country’s largest lenders recently said that the bank still did not have a timeline for when access could be expected.
"We have begun the rollout of Mythos 5 to organizations outside the United States. We continue to coordinate with the U.S. government to expand access to the broader set of domestic and international partners," an Anthropic spokesperson said.
Rees and Rohit Dhawan, head of AI at Lloyds Banking Group and another finance ministry appointee, have put together a set of recommendations aimed at shaping AI policy and regulation to accelerate adoption across the financial sector. Those proposals were published as part of the government's financial services AI adoption plan and include calls for regulators to review the growing use of AI chatbots for delivering financial advice to consumers.
The recommendations were released ahead of an upcoming annual speech by the finance minister and were accompanied by a government statement saying it has "set out a serious plan for AI sovereignty, backing British companies to compete and win."
Regulators have previously cautioned that concentration among a small number of critical technology providers can create financial stability and operational risks. To address both transparency and resilience concerns, Dhawan suggested that the UK could bring certain AI firms within the remit of financial regulators by designating them as "critical" providers to the finance sector - a move paralleling recent designations that included four U.S. cloud providers.
Alongside efforts to develop domestic capability, Rees recommended that the UK broaden its options by cultivating relationships with AI firms based outside the United States, citing providers in countries such as China and France as potential partners. The aim is to reduce single-source dependence and increase choices for banks procuring AI tools and services.
Industry and policy proposals now emphasize a twofold approach: invest in homegrown AI infrastructure, models and skills while simultaneously enlarging the pool of external suppliers and strengthening regulatory oversight to address concentration and operational risk.
Implications for markets and sectors
- Banking and financial services - AI access affects cybersecurity, operational resilience and the delivery of automated customer advice.
- Technology providers and cloud services - concentration and designation as "critical" providers could change regulatory obligations and scrutiny.
- National policy and skills - demand for domestic AI infrastructure and talent may influence public investment and industry hiring.