Written by Silvia Pavelli
Britain’s central bank is testing what happens when machines start making the same bad decision at the same time. That is a much more serious story than chatbot hype.
The smartest thing in the Bank of England’s new AI warning is that it sounds almost boring. According to PYMNTS’ report, deputy governor Sarah Breeden says the bank is running scenario analysis on how AI could hit markets and financial stability. In her letter, the point is not that AI will become a supervillain. It is that lots of firms could end up using similar models, similar data and similar incentives.
That is the nightmare: not machine genius, but machine conformity at machine speed. If enough trading systems start reading the world the same way, then “diversified” markets become an illusion. The Bank is explicitly studying whether AI agents could amplify stress through correlated behavior and herding. That is central banking language for a digital stampede.
The Treasury Committee was right to push for AI-specific stress testing, because finance has always been less fragile than synchronized. Humans panic in clusters already. Software just makes the cluster faster, cheaper and more scalable.
This is why the next phase of AI regulation should spend less time performing morality theater around chatbots and more time asking who is training models, on what data, with what objectives, and how many institutions are quietly buying the same black box. The financial system does not need artificial general intelligence to break. It only needs artificial consensus.
