Are you a client?
Sign in to view the full news archive.
The UK government's response to the Public Accounts Committee's (PAC) March 2025 report on AI adoption reveals encouraging commitments but also troubling patterns. While DSIT acknowledges PAC's criticisms and promises concrete action, the response also demonstrates how little has fundamentally changed in the government's approach to digital transformation. The real test will be whether renewed commitments translate into measurable progress.
The PAC's conclusions were not surprising (see AI in Government: Promise vs preparedness amid civil service cuts | TechMarketView). Out-of-date legacy technology remains a barrier, transparency standards are moving at a glacial pace, and there's a persistent skills shortage. Most tellingly, the committee found "few examples of successful at-scale adoption across government" – an indictment of multiple AI pilots, over several years, that have failed to graduate beyond proof-of-concept.
DSIT appears to accept these criticisms rather than dismissing them. The commitment to prioritise legacy technology remediation through Spending Review processes suggests recognition that you can't build AI capabilities on creaking infrastructure. Similarly, the acknowledgement that algorithmic transparency standards need proper rollout shows awareness that compliance theatre won’t cut it with the UK public.
However, it doesn’t take much scratching beneath the surface to find familiar patterns. The government continues to grapple with how to accelerate AI adoption across the public sector. One of the responses on this occasion is to establish a Public Sector AI Adoption programme to systematically gather insights from pilots, identify scalable AI products, and build cross-government capability to move initiatives beyond proof-of-concept stage. It feels like the latest in a long line of attempts to find the right organisational approach to accelerating AI adoption and realising its potential (Plans and progress of UK government AI adoption | TechMarketView).
The skills challenge remains particularly intractable. The government's commitment to having one in ten civil servants working in digital roles by 2030 sounds impressive until you consider they're simultaneously trying to reduce administration costs by 16% and cut consultant spend by £700m annually. There is bound to be a lag before the widespread—and impactful—implementation of AI brings the desired savings. As such, cutting costs and accelerating AI progress in parallel will not be easy.
Perhaps most revealing is the government's approach to AI procurement. The promise of a new framework that will "get the best from all suppliers" while ensuring "opportunities are available for small suppliers" sounds remarkably similar to every digital procurement reform of the past decade. The establishment of a Digital Commercial Centre of Excellence feels like déjà vu for anyone who has followed repeated attempts to fix the government's relationship with technology suppliers (*UKHotViewsExtra* Spending Review 2025: Digital transformation takes centre stage | TechMarketView).
The government's target dates tell their own story. Most commitments stretch to late 2025 or early 2026 – that’s if they are met. The pledge to strengthen spend controls and improve algorithmic transparency won't be truly tested in time to respond to the current urgency. What's missing is any sense of the pace at which AI is advancing in the private sector. While the government tackles the barriers to progression inherent in the UK public sector (see *NEW RESEARCH* A Blueprint for a Modern Digital Government | TechMarketView), it will struggle to leverage the potential of AI and the much-needed productivity benefits it could deliver.
Posted by: Georgina O'Toole at 09:59
Tags:
AI
public+sector