This page covers the Department for Science, Innovation and Technology (DSIT), the UK ministerial home of AI policy since 2023, and the Centre for Data Ethics and Innovation (CDEI), an expert advisory body now operating within DSIT under the Responsible Technology Adoption Unit (RTA) branding. DSIT is the UK government's lead department for AI, data, digital, and emerging-technology policy and the parent of the UK AI Safety Institute.
Department for Science, Innovation and Technology (DSIT)
| Field | Value |
|---|---|
| Type | UK ministerial department |
| Created | February 2023 (carved out of the former Department for Business, Energy and Industrial Strategy / DCMS) |
| Secretary of State | Peter Kyle (from July 2024); Liz Kendall (in post by January 2026) |
| Abolished | July 2026; functions redistributed, with the former Department for Business and Trade becoming the Department for Business, Innovation, Science and Trade (BIST) |
| Known for | UK AI policy lead; parent of the UK AI Safety Institute and CDEI/RTA |
Mandate and AI portfolio
DSIT is the UK government's lead department for AI, data, digital, and emerging-technology policy. Its AI portfolio includes hosting the UK AI Safety Institute (established November 2023; renamed UK AI Security Institute February 2025); publishing the UK AI regulatory approach, described as a "pro-innovation" framework of sectoral regulator guidance rather than a cross-cutting AI Act; ownership of the delayed UK AI Bill (see UK AI Bill — Status and Delay (Source Summary)); coordination of the Bletchley (2023), Seoul (2024), and follow-on AI summits (see The Bletchley Declaration (AI Safety Summit, 1–2 November 2023), Frontier AI Safety Commitments (Seoul, 2024)); and hosting CDEI / RTA.
July 2026 abolition
UK Prime Minister Andy Burnham's abolition of DSIT became public on July 21, 2026, with AI minister Kanishka Narayan elevated to cabinet in the same reorganization (Source: bbc.com). The July 20, 2026 appointment made Narayan the UK's first dedicated AI minister (Source: techieray.substack.com). The destination of DSIT's AI portfolio — including sponsorship of the AI Security Institute and ownership of the delayed UK AI Bill — was not detailed in the initial reporting. Parliament's Joint Committee on Human Rights later recorded that DSIT's functions were redistributed, that the former Department for Business and Trade became the Department for Business, Innovation, Science and Trade (BIST), that Narayan became Minister of State (Minister for Artificial Intelligence) jointly in the Cabinet Office and BIST, that the AI Security Institute moved to the Cabinet Office, and that the Government Digital Service became part of the Department for Digital, Culture, Media and Sport, citing a Written Ministerial Statement of 21 July 2026 and the Government's machinery-of-government fact sheet of 22 July 2026 (Human Rights and the Regulation of AI (Joint Committee on Human Rights, HC 160, September 2026)). Neither source held here assigns ownership of the AI Bill.
Regulatory posture
The UK has deliberately not pursued a single horizontal AI Act in the style of the EU AI Act. DSIT's approach instead rests on four elements.
It is principles-based: a 2023 AI regulation white paper established five principles (safety, transparency, fairness, accountability, contestability) that sectoral regulators apply in their domains. It is sectoral regulator-led: the Information Commissioner's Office (ICO), the CMA, the FCA, Ofcom, the MHRA, and others each issue AI guidance within their statutory remits. It relies on voluntary AISI access: the UK AISI operates pre-deployment model access arrangements with frontier labs on a voluntary basis, with no statutory authority. And it has so far deferred statutory codification: the UK AI Bill has been pushed to post-May 2026, with DSIT citing industry consultation, alignment with US posture, and preserving UK competitiveness as reasons.
In written evidence to the Joint Committee on Human Rights, DSIT stated that "the UK's domestic legal framework is consistent with the principles set out in" the Council of Europe Framework Convention on AI, described its approach as "targeted, risk-led," and said it favoured regulation "at the point of use by our existing expert regulators," while committing to "work with regulators to boost their capabilities." It described the Algorithmic Transparency Recording Standard as "a transparency and communication standard, not an assurance or audit mechanism." The committee concluded that the resulting framework is "fragmented and difficult to navigate" and applies mainly at the point of deployment, and recommended a cross-sector AI Bill and a single statutory AI oversight body (Human Rights and the Regulation of AI (Joint Committee on Human Rights, HC 160, September 2026)).
The UK Data (Use and Access) Act 2025 is the only AI-adjacent UK statute of the current period. It reforms automated decision-making and reconstitutes the ICO as the Information Commission, but defers frontier-model regulation. The UK's sectoral approach is explicitly positioned against the EU's horizontal regulation in the EU AI Act.
Centre for Data Ethics and Innovation (CDEI) / Responsible Technology Adoption Unit
| Field | Value |
|---|---|
| Type | Expert body / government unit |
| Founded | 2018 (as CDEI, within the Cabinet Office and later DCMS) |
| Current form | Responsible Technology Adoption Unit (RTA) within DSIT |
| Known for | AI assurance roadmap; algorithmic transparency standard; AI portfolio of applied-policy publications |
CDEI was established in 2018 as an expert advisory body to the UK government on data and AI ethics. Its core outputs included the AI Assurance Roadmap (2021), which argued for a mixed ecosystem of standards, audits, and certifications; the Algorithmic Transparency Recording Standard (ATRS), a framework for public-sector algorithm disclosure; and applied research on public attitudes, procurement, and regulator capacity.
Around 2023–2024, CDEI was rebranded and restructured as the Responsible Technology Adoption Unit (RTA) within DSIT, narrowing its mandate to applied public-sector AI adoption and reducing its arms-length advisory positioning.
The Joint Committee on Human Rights attributes the ATRS in its current form to the Government Digital Service. It is mandatory for all government departments and for arm's-length bodies that deliver public or frontline services or interact directly with the public, but not for local authorities or police forces, and it has no statutory basis. Several witnesses raised concerns about gaps in the published records; barrister Louise Hooper said it "has been mandated but we know it is not being used properly" and "Not all algorithmic use is on there." The committee concluded that the ATRS "provides some transparency concerning the use of AI systems in the public sector, but is limited in its coverage," and the Law Society proposed extending it to the private sector (Human Rights and the Regulation of AI (Joint Committee on Human Rights, HC 160, September 2026)).
People
Peter Kyle became Secretary of State for DSIT in July 2024; Liz Kendall MP held the post by January 2026 (Human Rights and the Regulation of AI (Joint Committee on Human Rights, HC 160, September 2026)). The DSIT Permanent Secretary and CTO roles rotate, and CDEI/RTA leadership has rotated multiple times during the restructuring.
Related sources and outputs
- Data (Use and Access) Act 2025 — Source Summary — the UK's principal current AI-adjacent statute; reconstitutes the ICO
- UK AI Bill — Status and Delay (Source Summary) — status of the delayed AI Bill
- UK AI Safety Institute (AI Security Institute) — DSIT-housed evaluation agency
- UK AISI — Advanced AI Evaluations May Update, UK AISI Frontier AI Trends Report 2025 — AISI outputs
Relationships
- related: UK AI Safety Institute (AI Security Institute) — DSIT-housed
- related: Peter Kyle — Secretary of State from July 2024
- related: Human Rights and the Regulation of AI (Joint Committee on Human Rights, HC 160, September 2026) — records DSIT's abolition, its evidence on the UK regulatory approach, and the ATRS's limited coverage; Joint Committee on Human Rights (UK Parliament)
- related: UK Information Commissioner's Office (ICO) — principal AI-adjacent regulator; reconstituted by DUAA
- related: Data (Use and Access) Act 2025 — Source Summary, UK AI Bill — Status and Delay (Source Summary)
- related: The Bletchley Declaration (AI Safety Summit, 1–2 November 2023), Frontier AI Safety Commitments (Seoul, 2024) — UK-convened international outcomes
- contradicts (framing): EU AI Act (Regulation 2024/1689) — the UK's sectoral approach is explicitly positioned against EU horizontal regulation