The Bank of England is allowing an AI model to read interest rate decisions first so it can gauge how markets will react.

Governor Andrew Bailey revealed that the Monetary Policy Committee (MPC) is submitting the minutes of its rate-setting meetings to a large language model (LLM) to see how it summarises them.

It is seen as useful as the Bank tries to predict how the language in its minutes and quarterly monetary policy reports will be interpreted by traders in the age of AI.

Bailey said: ‘We are very conscious that LLMs are now being used extensively.

‘I have very mixed views on this. It is helpful but you think “hmmm”, you know, where are we going with this?

‘But it’s been very helpful in the sense of just giving us a sort of summary view.’

Bailey has previously expressed disquiet about the wider impact of AI.

Earlier this year, the Bank’s governor warned of a ‘triple whammy’ of threats posed by AI, including soaring stock valuations, growing cyber-attack risks and automated trading.

He has also raised fears that AI could ‘destroy jobs’ faster than anyone expects.