UK banks’ inability to access Anthropic’s powerful AI model Mythos emphasises the urgent need for Britain to build its own AI capabilities and adopt a more strategic approach to the technology, a government-appointed banking sector adviser said.
Harriet Rees, the Starling Bank chief information officer who was appointed “AI champion” by the finance ministry in January, said Britain needed to build AI infrastructure and models while developing skills to ensure Britain is not fully reliant on U.S. tech.
“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 in an interview with Reuters.
U.S.-based Anthropic released Mythos in April to a select group that included U.S. lender JPMorgan.
STRICT U.S. RESTRICTIONS ON MYTHOS ACCESS
Banks want Mythos because it is said to be the most advanced AI model for identifying cybersecurity vulnerabilities, helping them to patch weaknesses more quickly and strengthen defences.
In Britain, only a very small number of banks — mainly the UK operations of U.S. lenders — were granted access, Rees said.
Big British lenders still do not have a timeline for when they expect to gain access to Mythos, the CEO of one of Britain’s largest banks said last week.
The U.S. administration introduced strict restrictions on who can access Mythos but at the end of June Anthropic said it had permission to release its AI model to trusted U.S. organisations.
An Anthropic spokesperson said: “We have begun the rollout of Mythos 5 to organisations 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.”
Bank of England Governor Andrew Bailey used his Mansion House speech on Tuesday evening to renew his call for the U.S. to take a more cooperative approach to tackling risks from frontier AI models.
“This needs to be coordinated internationally because no country can seal itself off from the cross-border nature of systems that are prevalent today,” he said.
PACKAGE OF RECOMMENDATIONS
Rees and fellow ministry appointee Rohit Dhawan, Lloyds Banking Group’s head of AI, have developed a package of recommendations on AI policy and regulation that aim to boost use of the technology.
The proposals, which also urge regulators to review the growing use of AI chatbots to deliver financial advice to consumers, were published on Tuesday as part of the government’s financial services AI adoption plan ahead of finance minister Rachel Reeves’ annual Mansion House speech.
The UK government said it would work with regulators and industry on the next steps.
In a statement, Reeves said she has “set out a serious plan for AI sovereignty, backing British companies to compete and win”, adding that the AI Adoption Plan is central to achieving this across the financial services sector.
Regulators have long warned that concentration among critical technology providers could pose financial stability and operational risks.
Alongside work to boost domestic capability, the UK should increase its options by building relationships with AI companies outside the U.S., such as those in China and France, Rees said.
To improve transparency and strengthen the resilience of the AI companies on which banks already depend, Dhawan added that the UK could bring them within the remit of financial regulators by designating them as “critical” providers to the finance sector, joining four U.S. cloud providers announced on Friday.
(Reporting by Phoebe Seers）
