Artificial intelligence could replace thousands of civil service jobs in fraud detection, a Cabinet Office minister has said.

Alex Burghart also said the technology could become the “institutional memory” of Government departments with a high staff turnover.

The minister was discussing the ongoing work to look at where AI can be used to improve efficiency, detect fraud, reduce error and increase productivity in Whitehall.

You can certainly envisage a future in which you have a smaller civil service than you have today

Asked what this means for the civil service workforce, he told a Centre for Policy Studies event: “We may not need to employ thousands of people to do fraud detection in the future.

“I hope we don’t. I hope that that’s something that we can make infinitely easier and cheaper for the British public…

“As we master this technology, you can certainly envisage a future in which you have a smaller civil service than you have today.”

The Tory MP for Brentwood and Ongar also spoke of the development of an AI red box, in which ministers receive important papers.

“What it does is it can read documents that go into your red box, it can summarise them, it can highlight connections between papers, connections between previous papers.

“And over time, as we fine-tune this model, it will become, I believe, the institutional memory of the department.”

Staff in the Cabinet Office “don’t always stay that long,” he said, meaning the loss of people who remember “things that happened three, four or five years ago”.

“But with an effective AI red box, that won’t be a problem,” he added. “We will be able to retain the experiences of previous policies and previous successes.”

The digital ministerial briefcase is being used by several ministers while it is being fine-tuned, and once ready will be offered to all colleagues, he said.

The hope is that before long, the technology can also be used to summarise MPs’ statements in the Commons and spare the time officials spend on administrative work.

Mr Burghart said: “We’re building these systems right now that we hope will enable us to go to the Treasury for the Budget in spring and say, ‘we are starting to prove the potential of these systems in Whitehall and help us go further’.”

The panel discussion in central London came against the backdrop of intense political focus on the Post Office scandal, with the errors of the Horizon software doing nothing to boost public trust in big tech systems.

Prime Minister Rishi Sunak is eager for the UK to be a key player in AI regulation, having hosted world leaders and industry figures at Bletchley Park for the world’s first AI Safety Summit in November.