UK’s FCA names eight banks for AI testing

The Financial Conduct Authority has picked 8 companies, among them Barclays, Experian, Lloyds Banking Group’s Scottish Widows and UBS, for a programme that volition proceedings AI applications successful unrecorded conditions.

The regulator is moving the AI Live Testing strategy with Advai, a London-based institution focused connected automated AI assurance.

The programme is intended to assistance participating firms analyse issues specified arsenic hazard controls and unrecorded oversight arsenic they hole to usage AI successful services affecting consumers and fiscal markets.

The applications showed however rapidly the tract is changing, covering a wide premix of AI approaches. These ranged from agentic AI and tiny connection models to newer systems specified arsenic neurosymbolic AI.

In the 2nd cohort, firms are assessing some user and business-to-business uses. These see AI-supported concern guidance, recognition people accusation for consumers, agentic payments, anti-money laundering checks and Know Your Customer functions.

Applications for the 2nd circular opened successful January 2026 and investigating started successful April.

The enactment is scheduled to decorativeness by the extremity of the year, followed by an appraisal study successful the archetypal 4th of 2027.

Alongside Barclays, Lloyds and UBS, the different firms selected are Experian Plc, Aereve, Coadjute, GoCardless and Palindrome. The FCA said the trials would instrumentality spot successful a controlled live-market mounting with existent customers involved.

An earlier group, which included NatWest Group, Monzo Bank and Scottish Widows, tested AI applications for astir six months.

The FCA said applications to its Regulatory Sandbox and Innovation Pathways were up 49% connected the erstwhile year. The study besides recovered that enactment successful the fintech marketplace broadly aligned with request for the regulator’s innovation services, particularly successful areas specified arsenic AI.

The regulator said it would people a study aboriginal successful 2026 mounting retired examples of bully and mediocre signifier successful the usage of AI successful fiscal services.

Separately, UK regulators, the Bank of England, the authorities and the National Cyber Security Centre are successful adjacent discussions implicit imaginable risks linked to Anthropic’s unreleased AI exemplary Mythos, reported Bloomberg citing sources.

Earlier this month, US Treasury Secretary Scott Bessent called Wall Street executives to an urgent gathering to guarantee they understood imaginable aboriginal risks from the technology.

FCA main data, accusation and quality serviceman Jessica Rusu commented: “We’re continuing to collaborate with firms to enactment the harmless and liable improvement of AI successful UK fiscal markets.

“With tailored enactment from the FCA and Advai, the inaugural reflects our committedness to supporting the gait of alteration successful AI, whilst demonstrating however regulators and manufacture tin enactment unneurotic to harness innovation responsibly.”