UK Banks Face Limited Access to Advanced AI Models
The restricted access for UK banks to Anthropic's advanced artificial intelligence model, Mythos, highlights a pressing requirement for Britain to develop its own AI infrastructure and implement a more strategic technological framework, according to a government-appointed adviser to the banking sector.
Harriet Rees, Chief Information Officer at Starling Bank and an "AI champion" appointed by the finance ministry in January, stated that Britain must build AI infrastructure and models, alongside developing necessary skills. This approach aims to prevent complete dependence on technology providers based in the United States. Rees emphasized the urgency of this initiative, noting that strategic thinking is crucial to safeguard the UK's leading position, with limited time available for action.
The Mythos Advantage and Access Disparity
Anthropic, a U.S.-based company, launched Mythos in April, initially granting access to a select group that included American financial institutions like JPMorgan. Banks are keen to utilize Mythos due to its reputation as a highly advanced AI model for identifying cybersecurity vulnerabilities, which can help them address weaknesses more rapidly and enhance their defensive capabilities.
In Britain, only a small number of banks, primarily the UK operations of U.S. lenders, have been granted access to Mythos, according to Rees. Major British lenders currently lack a timeline for when they might expect to gain access to the model, as confirmed by the CEO of one of the UK's largest banks last week. An Anthropic spokesperson indicated that the rollout of Mythos 5 to organizations outside the United States has commenced, with ongoing coordination with the U.S. government to broaden access for both domestic and international partners.
Recommendations for AI Policy and Regulation
Rees, in collaboration with Rohit Dhawan, head of AI at Lloyds Banking Group and another ministry appointee, has formulated a comprehensive set of recommendations concerning AI policy and regulation. These proposals are designed to accelerate the adoption and effective use of AI technology within the financial sector.
The recommendations, which also advocate for regulators to examine the increasing deployment of AI chatbots for providing financial advice to consumers, were released on Tuesday. This release coincided with the government's financial services AI adoption plan, preceding Finance Minister Rachel Reeves' annual Mansion House speech. The UK government has confirmed its acceptance of the recommendations directed at it, pledging to collaborate with regulators and industry stakeholders on subsequent steps. Reeves articulated a commitment to "AI sovereignty," supporting British companies to compete and succeed, with the AI Adoption Plan being central to this objective across the financial services sector.
Addressing Concentration Risks and Future Strategies
Regulators have consistently cautioned that a high concentration among critical technology providers could introduce significant financial stability and operational risks. Beyond developing domestic capabilities, Rees suggested that the UK should diversify its options by fostering relationships with AI firms located outside the U.S., specifically mentioning those in China and France.
To enhance transparency and bolster the resilience of the AI firms that banks currently rely on, Dhawan proposed that the UK could extend the purview of financial regulators to include these entities. This would involve designating them as "critical" providers to the finance sector, aligning them with four U.S. cloud providers that were similarly designated recently.
