- Bunce
- Posts
- 4 ways the Government has already implemented its “AI Action Plan”
4 ways the Government has already implemented its “AI Action Plan”
+ the tech they’ll need to roll out the rest of it
Joe Manktelow-Pimm
April 25, 2025
£14 billion in private investment.
13,250 new jobs (pending).
And one supercomputer powerful enough to "play itself at chess half a million times a second."
Here’s how the U.K. government is already implementing its lofty AI blueprint.
Slimming down the Civil Service
"No person's substantive time should be spent on a task where digital or AI can do it better, quicker and to the same high quality and standard."
A 12-week "AI accelerator programme" is already training 25 government staffers on how to build AI tools to get work done faster and better.
And the long-term vision is to have one in ten civil servants working in digital or data roles by 2030.
AI project leads will be appointed for specific government ‘missions’ to identify the ‘quick wins’ where AI can be leveraged to eliminate busywork.
"more than £45bn can be saved by greater use of digital methods in Whitehall, even before AI is deployed"
Hitting that infamous housing target
The Department for Science, Innovation and Technology (DSIT) has developed an AI tool called "Extract" that promises to clear the backlogs that hamper the planning process.
"Sluggish processes and outdated documents had made it 'nearly impossible' for local authorities to take informed decisions quickly"
It aims to digitise mountains of maps and notes into machine-readable, shareable data.
Apparently, it’s already performing tasks that typically take human planners one to two hours in just 40 seconds
"Technology like this could be a vital step towards councils meeting targets to help build the 1.5 million new homes the country needs, all while updating and improving the planning system for the future,"
Extract is currently being safety tested, with rollout possible as early as this year.
Speedier Public Services
AI has already become "deeply embedded in benefits processes", helping process claims and detect fraud according to the DWP
Though not without controversy.
Their "online medical matching” processes claims for employment and support allowance (ESA) by matching claimants' reported symptoms to conditions in the DWP's incapacity reference guide.
Initially used covertly from 2020-2024, the tool only achieved 35% accuracy.
It was mistranslating terms like "chronic fatigue" as "chronic renal failure", which is… not good.
But the DWP claims an updated version has reached 87% accuracy.
Another tool called "white mail" reads benefit claimants' letters to help prioritise vulnerable cases.
"We have started to build generative AI chatbots to change how people interact with the state, AI helpers to put an end to the mindless hours we spend on hold waiting for someone to pick up the phone, and tools to help get the views of citizens on policy proposals much more quickly – but AI can help with so much more,"
This recent study suggested that over 8 million hours could be saved every year by DWP and HMRC alone with the current AI tools
Now imagine that power compounded over 10 years of innovation.
"There is no reason people shouldn't expect the same experience from public services, as they get from the most innovative businesses."
This is just the start, with the government’s "Scan > Pilot > Scale" approach aiming to weave AI into "tens or hundreds of millions of citizen interactions across the UK."
The UK government also plans to roll out AI sensors that can monitor vulnerable people in their homes as "a major part of the plan for social care."
This system is designed to detect if an elderly person has fallen or missed a meal, automatically sending alerts to caregivers or family members.
The hope is that this will free up thousands of hours of precious care worker time in a sector plagued by staffing shortages.
Which would be lovely.
How the hell are they going to do all this
TL;DR: private sector money. Lots of it.
Cash injections (and job generation)
The government has secured commitments of £14 billion from three major tech companies (Vantage Data Centres, Nscale, and Kyndryl).
They’re going to build ‘data campuses’ which should create 13,250 jobs across the UK.
Plus, you’ve potential future investments from projects like Stargate, the $500 billion US behemoth, which are reportedly weighing up the opportunity.
Supercharging Blighty’s computing power
The government has committed to increasing public computing capacity twentyfold by 2030.
This begins with the construction of a new supercomputer with enough AI power to "play itself at chess half a million times a second."
All of this is going to require a lot of energy…
…so the government is establishing AI Growth Zones, with the first one planned for Culham, Oxfordshire.
They’ll look to land a private-sector partner to develop "one of the UK's largest AI data centres," starting with 100 MW of capacity and plans to scale up to 500 MW.
For context, that’s similar to an average power plant. Just for UK-centric AI.
Landing AI talent (which has historically fled to the States)
But the hardware alone isn’t enough – we need people who can build and direct these systems.
So the government wants to train "tens of thousands of additional AI professionals” over the next five years.
They’re planning an undergraduate and masters AI scholarship program "on the scale of Rhodes, Marshall, or Fulbright" for students to study in the UK, coupled with “headhunting capability on a par with top AI firms" to bring the best to the UK.
Social Care