AI takes a seat at the Cabinet table as the UK abolishes DSIT
The UK has moved artificial intelligence to the centre of government. In the machinery-of-government changes that followed Andy Burnham becoming prime minister on 20 July, the Department for Science, Innovation and Technology (DSIT) — created only in 2023 — was abolished, and responsibility for AI was handed to the Cabinet Office. Alongside it, Kanishka Narayan was appointed Minister of State for Artificial Intelligence, the first holder of an AI brief to attend Cabinet.
We have not covered the reshuffle until now, and we should have: it reshapes the institutions behind almost every UK AI story we write.
Where everything went
DSIT’s remit was split three ways. Science and innovation moved into an expanded Department for Business, Innovation, Science and Trade. Digital functions, including the Government Digital Service, went to a strengthened Department for Digital, Culture, Media and Sport. And AI — strategy, public-sector adoption, and crucially the AI Security Institute — moved into the Cabinet Office, the department that sits closest to the prime minister. Narayan holds his post jointly across the Cabinet Office and the new business department.
That last move is the substantive one. The AI Security Institute, whose work testing frontier models and agents has given the UK a genuine voice on AI safety, now reports up through the centre of government rather than a line department. The £500M Sovereign AI programme, which has been taking equity stakes and handing out compute to British companies, sits under the new arrangements too.
The signal, and the risk
The symbolism is hard to miss. Putting AI in the Cabinet Office, with a minister at the Cabinet table, says the government intends to treat it as a cross-cutting national priority rather than one sector among many in a science department. For an ecosystem that has spent the past year watching talent and compute pulled toward better-funded American labs, a state that wants AI run from the centre is, on its face, a statement of seriousness.
The risk is the reorganisation itself. Machinery-of-government changes are expensive and disruptive, and the Institute for Government warned that dismantling DSIT risks “losing focus, expertise and coordination” that a dedicated department had begun to build. Splitting science from AI, and AI from digital, also cuts across a field where those things are deeply entangled — the same chips, the same data infrastructure, the same people. A minister at the Cabinet table is worth little if the officials and continuity beneath moved out of the door in the process.
What it changes for us
Practically, it changes the map. The bodies we reference constantly — the Sovereign AI Unit, the AI Security Institute, the AI Opportunities Action Plan — now answer to a different part of Whitehall, and some of our earlier coverage refers to DSIT as though it still exists. We are updating those references, and this is the piece we will point back to. Whether elevating AI to the Cabinet Office proves to be more than a reshuffle will show in the ordinary things: whether the Sovereign AI fund keeps deploying, whether AISI keeps its independence and expertise, and whether the government can actually buy from the companies it says it wants to back. Those are the tests, and we will keep watching them.
Read the original story on GOV.UK .
Previous coverage
- UK's Sovereign AI fund names its first companies, from drug discovery to chips 4 August 2026
- UK's Sovereign AI fund now backs 12 companies, and wants to be their customer 20 August 2026
- UK AI Security Institute finds agents going off-script in cyber tests 13 August 2026
- UK AI Security Institute launches Control Red Team, finds flaws in every safety monitor it tested at Anthropic and DeepMind 24 July 2026
- UK opens its first AI 'growth lab', a regulatory sandbox for legal services 10 August 2026