Bank of England plans to ease capital rules despite AI stability fears

. UK edition

The Bank of England in London with City buildings in the background
The Bank of England’s financial policy committee is to review whether its plans ‘would leave any financial stability gaps’. Photograph: Amer Ghazzal/Shutterstock

Central bank’s financial policy committee members voice concern on trimming big lenders’ financial buffers

The Bank of England is planning to loosen capital requirements for major UK lenders, even as policymakers expressed concern about the threat to financial stability from rapid AI developments and debt-fuelled stock investments.

The central bank said on Tuesday it was looking to remove and loosen some rules introduced after the 2008 financial crisis that determine the size of the financial cushion required to absorb losses and protect consumers and taxpayers when things go wrong.

The Bank’s financial policy committee (FPC) said that included plans to scrap a longstanding buffer within the so-called leverage ratio, in a way that would primarily benefit the largest of the UK’s domestic-focused banks and building societies, including NatWest, Lloyds, Nationwide and Santander UK.

Current proposals, which will be put out for consultation, could slash those lenders’ leverage ratio by 20 basis points on average, helping give them a leg-up against international peers, and spur further lending that supports the wider UK economy.

However, some committee members have raised concerns that trimming those buffers could amplify current risks to the financial system.

A fresh wave of lending, for example, could increase the number of loans to investors, including hedge funds, who have already used a heavy amount of debt to buy company shares on the stock market.

Much of the debt-fuelled investments have been in AI-related stocks, whose valuations have soared in recent months.

“Some FPC members were concerned that the proposal might lead to an unwanted increase in market-based leverage, with implications for the resilience of core UK markets,” a report by the committee said.

The FPC is now embarking on a review that will “identify whether the proposal would leave any financial stability gaps that would need to be managed and whether this justified further adjustments to the policy package”.

That review, which will be completed by the end of September, will then influence the package of capital changes put forward for consultation in early 2027.

The Bank had already announced in December that it was lowering capital requirements related to risk-weighted assets by one percentage point to about 13%, reducing the amount lenders must hold in reserve.

The looser rules come after regulators came under pressure to do more to stimulate growth. Last summer the chancellor, Rachel Reeves, argued rules and red tape were a “boot on the neck” of businesses that risked “choking off” innovation across the UK.

The FPC raised further concerns about developments in AI, which had developed much more quickly than some experts had forecast. While frontier AI systems had the ability to boost productivity, it increased cyber risks significantly, meaning malicious actors could inflict shocks and outages at lower costs and at a greater scale.

That could hit banks and systemically important financial companies, putting the wider system at risk.

“Recent rapid advances in frontier AI capabilities have increased financial stability risks related to cyber and operational resilience,” the bank said.

The warning comes amid months of speculation and warnings over the impact of AI models such as Anthropic’s Mythos, which has only been rolled out to select vetted companies worldwide.

The Bank of England governor, Andrew Bailey, told journalists on Tuesday that access to Mythos “does tend to differ week by week”, referencing Donald Trump’s temporary ban on foreigners using the US company’s latest AI models.We are now, obviously, keen to work with Anthropic, and are talking to Anthropic about just how that will work,” he said.

He said the growing risks warranted closer scrutiny, and international coordination, to keep financial systems safe. “These models are a big step forward in terms of capabilities, and the threat issue is a really a major step forward … It is therefore important that we have a process by which we can address the issues that these models are turning up, and that there can be robust testing … It’s important that we work together internationally to do this because … the financial system it’s hugely interconnected,” he added.