A think tank has warned that government needs to "fundamentally rewire itself" to take advantage of the possibilities of artificial intelligence – and called for the creation of a new function backed with £1bn in funding to drive change.
A just-published report from Reform says that AI has huge potential to improve "stagnant productivity" in public services and transform the kinds of services provided.
However, the independent organisation argues that a government that is still catching up with the last generation of digital transformation will not be able to capitalise on the opportunities AI can offer, and needs to centralise its approach to drive progress.
Report authors Joe Hill and Sean Eke praise the new government's July decision to decision to move the Government Digital Service, Central Digital and Data Office, and AI Incubator from the Cabinet Office to the Department for Science, Innovation and Technology.
But they say that the machinery of government change, designed to make DSIT the "new digital centre of government", requires further strengthening.
Policy director Hill and researcher Eke call for the establishment of a Government Data and AI Service as a separate function within CDDO that would sit alongside GDS and have a remit of driving AI adoption across the public sector.
They argue that the GDAIS should incorporate the current Incubator for AI – known as i.AI – and should be led by a "government chief AI officer". The chief officer would be supported by programme directors for "priority" use cases – such as process automation and assessment streamlining – who would lead adoption across departments.
Hill and Eke say GDAIS should have a team of specialists who serve as the point of contact with each department and understand its business and pipeline of projects. They say the new operation should also collect evaluations of AI from across government and vouch for commonly used off-the-shelf AI products on behalf of central government.
They are calling on prime minister Keir Starmer and chancellor Rachel Reeves to use next month's Budget to create a new "AI transformation fund" and allocate £1bn to GDAIS between 2025-26 and 2027-28 to finance it.
The report authors' dozen-plus recommendations also include the establishment of a cross-government "specialist development scheme" for key AI roles, with pay frameworks set at levels capable of attracting candidates to join the civil service.
Hill and Eke said public service takeup of AI is "extremely limited" compared with the private sector and that government needs to start focusing on using AI in areas where there is already a strong evidence base and a "quick route" to adoption at scale.
"This transformation will not happen unless government fundamentally rewires itself to adopt the new tools much more quickly," they said.
"The public sector is still too risk averse to even test AI in many cases, and lacks the focus and investment required to see projects through to deployment at scale.
"Proving the government can successfully use AI, and reap widespread benefits from doing so, is essential to making the case for further adoption."
Hill and Eke said the government needs to reorient its approach to AI by introducing greater central leadership to drive adoption in public services and making funding available for rapid deployment to "sustain and scale up" successful pilots.
"Crucially, government needs to be willing to take smart risks in deploying AI," they said. "Government must assess use cases against the risks of continuing with the status quo – failing public services, declining accessibility, rising costs, and in some cases high levels of bias.
"This new approach requires accepting that transformed public services will not happen organically, and innovation needs to be led by the centre in partnership with frontline services. And it requires establishing parity between the risk of adopting AI with the risks of not doing so."