The Social Market Foundation (SMF) argues the fund, unveiled in March’s Spring Statement, lacks strategic clarity, leading to what it describes as “an inconsistent pattern of AI adoption across government.”
Freedom of Information requests submitted by the think tank reveal a wide variation in how departments are using AI, with some running pilots or deploying specialised tools, while others are still in the planning stages.
Funding, meanwhile, continues to be allocated through traditional Treasury bidding — a model the SMF calls “poorly suited to the agile, iterative nature of AI development.”
The Transformation Fund, unveiled by the Chancellor in March, is part of the Labour government’s wider efforts to harness AI, following earlier plans to ease regulatory hurdles for research and attract data centre investment.
The fund is intended to support digital technologies and AI projects aimed at improving public services, with an initial £42 million earmarked for 'Frontier AI Exemplars' — pilot schemes to demonstrate the tech’s potential within civil service workflows.
The SMF, however, contend that its FOI request found that some £150 million of the Transformation Fund is being used to pay for voluntary exit schemes for civil servants and £8 million for administrative tech in probation services, though it’s unclear what, if any, AI will be involved.
The think tank says departments are progressing at starkly different paces. Notably, the Department for Business and Trade reported 34 current or planned AI use cases as of April 2025, while the Treasury and Ministry of Housing, Communities & Local Government had just a handful that were largely in the planning phase.
Many AI initiatives remain at the pilot stage, the SMF said, with little visibility on whether they’ll be scaled or scrapped.
Sam Robinson, senior researcher and AI lead at the SMF, said: “As pressure mounts to meet civil service efficiency targets – including a 15% cut in running costs and 10,000 job losses – effective AI deployment is no longer a luxury.
“The Transformation Fund is a step forward, but in its current form, it risks becoming a general-purpose reform pot rather than a targeted lever. Too little of the money is being explicitly directed at scaling the most effective AI tools.”
The SMF previously identified some 8.12 million hours, it argues, could be saved per year by deflecting or streamlining caseload at the Department for Work and Pensions and HM Revenue & Customs by using AI and automation tools.
The think tank called for the government to shift funding allocation to a ‘fund teams, not programmes’ approach, whereby multidisciplinary teams in departments are given devolved budgets to experiment with AI solutions and “cut through bureaucratic inertia”.
“The government is right to aim high with AI, but a bold vision without clear execution risks leaving us with a patchwork of pilots and no real transformation. The Fund must empower civil servants to build and scale what works – and stop spending money on reinventing the wheel,” Robinson added.
RELATED STORIES
Spring Statement: UK pledges £13bn to boost infrastructure projects
UK fast-tracks AI infrastructure with new Growth Zones
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms
