Five UK banks reportedly restrict AI agents’ access to customer accounts

HSBC, Nationwide, NatWest, Barclays and Lloyds have moved to block or restrict AI agents from customer accounts, The Telegraph reported on November 8. Some restrictions specifically bar transactions without the customer’s direct involvement, drawing a line around software acting on someone’s money while responsibility for mistakes remains unclear.

The measures are pre-emptive: the newspaper said there were no signs yet of AI bots accessing the British accounts. Lloyds chief financial officer Will Chalmers likewise said his bank had not seen such activity. The immediate development is therefore a restriction on potential access, rather than a response to an identified wave of automated transactions.

Keeping the customer in the transaction

AI agents go beyond offering advice: they can carry out tasks for a user. In banking, that raises a different question from whether software can explain a balance or suggest a savings account. An agent allowed to transact would be acting on the customer’s behalf, making the scope of its authority—and who answers for an error—central concerns.

Banks are concerned that agents could act without authorization or mismanage customers’ money, with no clear rules assigning responsibility. The reported transaction bans at some institutions preserve direct customer involvement rather than letting an outside assistant complete the action alone. That is a narrower distinction than simply deciding whether AI can be useful for personal finance.

Existing UK authentication requirements also limit what an agent can do independently. Most payments must be approved through a mobile banking app, often using a fingerprint or face scan. Those checks keep a person involved at the payment stage, adding a practical constraint alongside the banks’ own access restrictions.

We do guard against that type of risk, and we take it very seriously.

Will Chalmers, Lloyds chief financial officer, quoted by The Telegraph

Better rates can threaten deposit profits

Customer safety is not the only concern described in the report. Banks also face a commercial risk if AI makes it easier for savers to move money from accounts paying little or no interest. A tool that helps an individual find a better return could, at scale, put pressure on lenders’ deposit profits.

McKinsey has warned that savers could shift up to £700 billion into savings products offering better returns on the advice of AI chatbots. Its scenario suggests that moving just 5% to 10% of bank balances this way could reduce lenders’ deposit profits by a fifth or more. Those are projections, not measured losses from agent activity.

A separate concern is speed and coordination: agents acting together could withdraw deposits from one institution at the same time, potentially triggering a bank run at machine speed. The feared outcome is not merely one assistant making a poor decision, but many automated decisions creating a much larger problem for a bank.

Rules are still taking shape

The Financial Conduct Authority is working with the financial sector and other regulators to develop rules for AI agents in finance. Sources close to the regulator told The Telegraph that individual banks must decide whether to permit access. They also said early research had shown agents could behave in unexpected ways.

The British restrictions follow comparable blocks at Citigroup, Bank of America, Wells Fargo and JPMorgan, according to the newspaper. That places the five UK institutions alongside major US lenders taking a restrictive approach to outside agents, while the rules governing their role in finance are still being developed.