TL;DR
Market chaos triggered by AI is one theme being discussed for the Bank of England’s next system-wide exploratory scenario (SWES), according to a Bloomberg report carried by Claims Journal on 21 October. The report relies on people familiar with the matter, and a second person said no decision has been taken. A third person said AI is not the only option on the table.
What the Bank is weighing
A SWES differs from the Bank’s routine stress tests. Rather than testing banks or insurers one at a time, it models how a hypothetical shock spreads across the whole financial system. The UK has run this kind of exercise only once before the current round.
That current round, now in its final stages and set to finish in 2027, looks at private markets, which have grown sharply since the financial crisis. It models a deep recession triggered by a severe global supply and geopolitical shock, and how that would choke the flow of credit through banks, investors and fund managers. AI already features at the edges of that scenario, which assumes AI valuations fall and the technology’s development stalls. According to the report, worries about AI and finance have broadened since the exercise began.
Three concerns are driving the discussion, Bloomberg says. Officials fear that trading firms using similar AI agents could all react to a shock in the same way, deepening a sell-off. They are watching for new models exposing weak cyber defences at financial firms. And they have long tracked the risk of an AI share-price bubble bursting and hitting the wider economy.
Bailey’s two-sided view
Governor Andrew Bailey has argued AI could lift Britain’s weak growth, while warning about how markets are pricing it. “Everybody’s priced to be a winner,” he said recently, adding that markets could overestimate future returns. On cyber risk, he said AI had become “a very big issue for us” from the spring onwards.
Looking forward
The report fits a run of finance-sector moves on AI risk, including the G7’s plans for AI cyber attack exercises and Bailey’s warning on AI-enabled fraud and trust. In our view, an AI-themed SWES would matter well beyond the largest banks. The exercise looks at how stress moves between institutions, so insurers, asset managers and the firms that supply them with AI-driven trading and risk tools could all find their assumptions examined. Firms relying on similar models from a small number of vendors may want to ask now how correlated their behaviour would be in a fast-moving market.