The £45 Billion Question: Why UK Government AI Planning Is Falling Short
The UK government has been broadcasting a massive, audacious target for years now: £45 billion in annual efficiency savings, all thanks to the magic of artificial intelligence and digital transformation. It’s an easy figure to throw around in speeches, the kind of number that sounds like a grand vision of a more productive, streamlined state. But while ministers have been focused on these stellar, long-term projections, they’ve largely ignored the mundane, difficult, and absolutely essential reality of how that change actually happens. This is a critical issue that also affects other government programmes, like the AI CV tool for jobseekers.
The National Audit Office (NAO) has finally arrived to pour some much-needed, sobering reality onto this bonfire of vanity. And the verdict? It’s not just "worryingly optimistic," as Parliament's Science, Innovation and Technology Committee previously warned; it’s bordering on negligent. The core problem, the NAO audit confirms, isn't just about the money. It's about the people.
Where the Maths Meets the Myths
The government's £45 billion claim was always a bit of a stretch. The Institute for Government has already pointed out the sheer, impossible magnitude of that number, especially considering how much of the public sector budget is locked into salary and infrastructure costs that aren’t easily "automated" away.
But the NAO’s latest findings highlight a more fundamental issue: the government is gambling on these savings without doing the basic groundwork. They aren’t just selling hype; they’re selling a vision without a roadmap. They are banking on AI sweeping through the civil service, making everyone instantly more productive, and freeing up massive amounts of budget, while completely failing to map what that actually looks like for the hundreds of thousands of people who work in the sector.
The Missing Human Element in Workforce Planning
If you want to realistically expect gains from AI, you have to talk about how work itself is going to change. And that’s the missing piece. The NAO is telling government leaders in no uncertain terms that they need to integrate major technical changes into their long-term workforce planning.
This isn’t just HR jargon; it is about the fundamental shape of the state. It means identifying exactly how roles, tasks, and grades will shift. What happens to the admin staff when the tax checking is automated? What skills does a civil servant need in 2030 that they don’t have in 2026? How many people will we actually need on the payroll?
Departments currently lack detailed documentation on how they derived these projected efficiencies. There is a glaring risk that these figures are being plucked out of thin air, with no consistent methodology. More concerning is the fear that departments haven't even begun to identify the full scope of opportunities—or the real, human costs—of this transition.
The Skills Deficit and Ethical Constraints
Even if the government did all the math tomorrow, they’d hit another wall immediately: they don't have the people to build the future they’re imagining. The public sector is currently facing a massive, structural deficit in the technical and digital skills required to actually deploy these AI systems. You can’t build an "AI-first" culture if you don't have the workforce that understands AI—both its potential and its deep limitations.
And let’s be honest, tech adoption in the public sector isn’t just a matter of flipping a switch. AI brings its own baggage. Even the NAO acknowledges that there are ethical concerns holding back deployment in some areas, simply because public trust and institutional accountability matter more than an efficiency spreadsheet. Departments are already looking at restricting AI use because, frankly, they know it can go wrong, and they aren't equipped to manage that risk.
A Reckoning with Reality
Cabinet Secretary Dame Antonia Romeo has already admitted that civil service workforce plans simply don't reflect how technology is about to reshape the organization. This isn't just a failure of a single department; it’s an institutional blind spot.
Government departments must merge their grandiose AI plans with their strategic workforce plans. They need to get serious about the training, the retraining, and the honest, difficult conversations about how this restructuring will play out.
The public deserves more than optimistic projections and magical thinking. They deserve a plan that accounts for the reality of the workforce and the complexity of moving to a new, digital way of operating. Until the government starts doing that, this £45 billion ambition is going to remain exactly what the critics have already labeled it: hype. And while they continue to chase that number, the actual work of building a sophisticated, modern, and capable public sector will continue to lag behind.
