BoE flags AI and war risks

Bank of England policymakers warned on Wednesday that artificial intelligence could become a fast-moving threat to financial stability as firms deepen its use, while the economic shock from the war involving Iran has already made the wider risk backdrop more volatile and harder to predict. The central bank’s Financial Policy Committee said advanced AI is not yet being used across the financial system on a scale that poses a systemic danger, but it judged that the risk could rise quickly as more institutions prepare broader deployment.

The warning marks a shift from watchful assessment to stronger caution. The committee said there is still little evidence that banks, insurers and market participants are relying on advanced generative or agentic AI for core decisions such as credit underwriting, insurance pricing or central trading activity. Even so, officials said firms’ intentions to expand adoption are growing and that the technology’s speed, opacity and potential to amplify common behaviour across markets mean the stability threat could build sharply rather than gradually.

That matters because the central bank is not looking at AI in isolation. It set the warning against a backdrop of rising geopolitical and macroeconomic stress after the war linked to Iran drove a fresh energy shock through global markets. The Bank said the conflict had inflicted a substantial negative supply shock on the world economy and increased the risk that already visible vulnerabilities could crystallise at the same time. Officials pointed to the danger of overlapping strain in government bond markets, private credit and richly valued US technology stocks, many of which have been central to the global enthusiasm around AI.

For Britain, the transmission channel is straightforward and uncomfortable. Disruption around the Strait of Hormuz has pushed up energy costs, with UK natural gas prices rising by more than 70% since the end of February, according to the Bank’s assessment cited by Reuters. Petrol prices are also higher, while household energy bills are due to rise from July under the regulated price cap. Higher energy costs have combined with tighter funding conditions to lift borrowing costs, including in the mortgage market, where two-year deals have become more expensive and a chunk of available products has been withdrawn.

Against that backdrop, AI-related exuberance looks more fragile. The Bank said valuations for large US technology groups with heavy AI exposure were already stretched and now appear even more so because data-centre economics are sensitive to power costs and supply-chain disruption. That does not amount to a prediction of a market correction, but it is a clear signal that policymakers see a link between the AI boom and broader financial conditions. When energy prices, rates and risk appetite move at the same time, sectors carrying premium valuations become more exposed to abrupt repricing.

The Bank’s concern over AI runs through four channels. One is the possibility that firms begin using advanced models in important financial decisions without fully understanding how those systems behave under stress. Another is greater use in markets, where shared models or signals could intensify herd behaviour and volatility. A third lies in operational concentration, as more institutions depend on a small group of common service providers. The fourth is cyber risk, including the use of AI by attackers and the possibility that more capable frontier systems create new vulnerabilities. Officials said agentic AI deserves particular attention because private incentives to deploy it may fail to capture wider systemic harm.

At the same time, the Bank has been careful not to cast AI only as a menace. It said the technology could deliver meaningful gains in productivity and growth across finance and the wider economy. That balancing act is important for regulators trying to avoid smothering innovation while still moving early enough to stop dangerous concentrations from forming. The committee said the regulatory approach should remain flexible and outcomes-focused, and the Bank is expanding its monitoring through surveys, supervisory intelligence, market feedback and work with domestic and international bodies.

There were also signs of resilience in the central bank’s broader judgement. It said households, businesses and banks in Britain remain in relatively strong shape, with indebtedness generally low by long-run standards. Yet that reassurance came with caveats. Officials noted that if markets continue to expect higher interest rates, a majority of mortgage borrowers could face higher repayments by the end of 2028. In private credit, the default of specialist lender Market Financial Solutions has already underlined how stress can surface in riskier corners of finance.