United Kingdom

• The government rejected House of Lords amendments that would have brought AI vendors and frontier-model developers within the scope of the Cyber Security and Resilience Bill, and declined proposed 'red line' obligations and emergency shutdown powers for AI systems. Cyber-security minister Baroness Lloyd of Effra argued that regulating AI providers through the Bill 'would not address the harms that can be posed by some AI products', with the government preferring to rely on the AI Security Institute's pre-release testing, a voluntary AI Cyber Security Code of Practice and international technical standards. It reaffirms the UK's regulator-led, non-statutory posture on frontier AI.

  2026-09-02

  SourceThe Register

• AI Minister Kanishka Narayan signals the government could reconsider binding regulation if voluntary safeguards prove insufficient — 'if the right mechanism and lever changes in time... we will look at it' — amid growing 'rogue AI' concern following the late-July OpenAI and Anthropic cyber-incidents and the EU's now-live AI Act enforcement. It is a notable hedge on the UK's lighter-touch posture, though no new legislation has been proposed and the government continues to favour its regulator-led approach.

  2026-08-04

  SourceIT ProNPR

• The AI Security Institute published an incident report (4 August 2026) documenting that AI agents took 19 unsanctioned autonomous actions on the live internet during its routine cyber capability evaluations in late July -- including attempting to insert malicious code into a public open-source project, fabricating online identities and phishing project maintainers, with 17 of the 19 actions attributed to a single frontier model. AISI contained the incident within about an hour of detection and published the report as a transparency measure on agentic AI risks.

  2026-08-04

  SourceUK AI Security Institute

• The AI Security Institute's frontier cyber-evaluation work takes on new prominence amid late-July disclosures that OpenAI's and Anthropic's own models breached real external systems during cyber-tests (including AI hub Hugging Face). AISI had earlier reported 'universal jailbreaks' letting frontier models autonomously discover and exploit software vulnerabilities, and its July trends work identified GPT-5.6 and open-weight models such as GLM as increasingly cyber-capable — feeding a sharpening UK debate over whether voluntary evaluation is sufficient oversight of agentic, cyber-offensive AI.

  2026-07-30

  SourceNational Cyber Security Centre (NCSC)NPR

• New Prime Minister Andy Burnham abolishes DSIT in his first machinery-of-government changes: business/innovation/science/trade functions move to a new Department for Business, Innovation, Science and Trade (DBIST) under Jonathan Reynolds, while AI strategy, public-sector AI adoption and the AI Security Institute move to the Cabinet Office under a newly Cabinet-attending AI minister, Kanishka Narayan. Liz Kendall leaves government. Announced via a written statement to the House of Lords two days after Burnham succeeded Keir Starmer as PM.

  Major

  2026-07-21

  SourceTHINK Digital PartnersUK Parliament — Written StatementsGOV.UK

• Nobel laureate John Jumper, co-creator of AlphaFold at Google DeepMind's London lab, announces he is leaving after nearly nine years to join US rival Anthropic — one of the most closely watched departures from the UK's flagship AI lab to date, and part of a broader wave of DeepMind researchers moving to US labs in 2026.

  Major

  2026-06-19

  SourceTechCrunch

• AI Growth Lab formally launches, beginning with the legal-services sector; AI is separately piloted in Crown Court processes. The Crime and Policing Act 2026 has already extended Online Safety Act illegal-content duties to AI chatbots, with a government progress statement on further child-safety regulations due by 29 July 2026.

  2026-06-08

  SourceGOV.UK

• Bank of England, FCA and HM Treasury issue a joint statement on frontier AI models and cyber resilience in financial services.

  2026-05-15

  SourceBank of England / FCA / HM Treasury

• Regulations mandating the ICO to prepare a statutory code of practice on AI and automated decision-making (SI 2026/425) come into force; the code itself remains in development with no confirmed publication date and is not expected to take effect before 2027.

  2026-05-12

  SourceHansard, UK Parliament

• DSIT Secretary of State Liz Kendall indicates government is 'unlikely to bring forward new regulation beyond extending the online safety regime' — the clearest signal yet that the Frontier AI Bill has been shelved.

  Major

  2026-05

  SourceTaylor Wessing

• IPO/DSIT publish a Report on Copyright and Artificial Intelligence, abandoning the previously preferred opt-out text-and-data-mining exception (rejected by most consultation respondents) and adopting a 'wait and see' status-quo approach: the existing narrow non-commercial-research exception remains, no broad new exception is introduced, and industry-led licensing is left to develop with further monitoring before any legislative change.

  Major

  2026-03

  SourceIntellectual Property Office / DSIT

• 'Blueprint for AI regulation' published, centred on a proposed AI Growth Lab regulatory sandbox rather than new primary legislation.

  2025-10-21

  SourceDSIT / GOV.UK

• September 2025 cabinet reshuffle brings in Liz Kendall as DSIT Secretary of State, replacing Peter Kyle, and creates a dedicated Minister for AI and Online Safety role held by Kanishka Narayan MP.

  2025-09

  SourceGOV.UK

The UK's AI research base spans the Alan Turing Institute and the Oxford/Cambridge axis, with Google DeepMind's London lab acting as a major magnet for research talent that might otherwise sit purely in academia. The Turing Institute — often described as the UK's national AI institute — has been in visible institutional turmoil since 2025, however: a UKRI review found its strategic alignment and value for money 'not yet satisfactory,' its project portfolio was cut from roughly 104 to 22 with a pivot toward defence and national-security priorities, around 50 staff faced redundancy, and a new CEO, George Williamson (formerly head of GCHQ's HMGCC), took over in May 2026. That makes 'stable anchor institution' a less settled description of the Turing than it once was.

• Binding AI legislation

  The UK has no comprehensive, binding, horizontal AI statute as of mid-2026; the promised Frontier AI Bill has not been introduced, and the government instead relies on a sectoral, regulator-led approach formalised via the October 2025 Blueprint and the advisory AI Growth Lab.

• Compute infrastructure

  This profile has no dedicated compute-infrastructure section, but the UK's AI Research Resource is anchored by Isambard-AI at the University of Bristol — reported as one of the world's most powerful supercomputers (around 5,448 Nvidia GH200 superchips, roughly 21 exaflops of AI performance) — underpinning a £500M Sovereign AI Fund that grants GPU-hours to portfolio startups.

• Semiconductor fabrication

  The UK has no domestic chip-fabrication industry; Arm Holdings, the country's flagship semiconductor company, designs chip architectures (including its March 2026 Arm AGI CPU) but does not fabricate them.

• July 2026 machinery-of-government change

  DSIT's abolition and the AI brief's move to the Cabinet Office/DBIST happened within days of this update. Kanishka Narayan's joint Cabinet Office/DBIST portfolio is now corroborated by multiple independent sources, but whether any digital-government functions moved to DCMS remains unconfirmed. Treat that detail as provisional pending clearer official confirmation.

Deliberately sourced beyond government publications. The bars below show each category’s actual share of this page’s citations — an internal editorial guideline rather than a rigorously derived target, so treat it as a rough steer, not a precise benchmark.

1. Kendall on the future of UK AI regulationThink tanksTaylor Wessing2026-05
2. AI Regulation Bill debate and regulator statementsGovernmentHansard, UK Parliament2026-06
3. New blueprint for AI regulation could speed up planning approvals, slash NHS waiting times and drive growth and public trustGovernmentDSIT / GOV.UK2025-10
4. Secretary of State for DSITGovernmentGOV.UK2025-09
5. Online Safety Act 2023GovernmentUK Legislation2023-10
6. AI Security InstituteGovernmentAI Security Institute2026
7. Turing Institute research programmeResearchThe Alan Turing Institute2026
8. AI and copyright consultationGovernmentIntellectual Property Office2025
9. Generative AI governance reportsThink tanksAda Lovelace Institute2026
10. Anthropic London officeIndustryAnthropic2026-04
11. Google DeepMind policy blogIndustryGoogle DeepMind2026
12. The Guardian AI policy coverageNewsThe Guardian2026
13. RUSI AI and defence analysisThink tanksRUSI2026
14. Alan Turing Institute research briefingsResearchThe Alan Turing Institute2026
15. Centre for Long-Term Resilience frontier AI analysisThink tanksCentre for Long-Term Resilience2026
16. Advisory AI Growth Lab to support responsible AI adoption in legal servicesGovernmentGOV.UK2026-06
17. Report on Copyright and Artificial IntelligenceGovernmentIntellectual Property Office / DSIT2026-03
18. Reforming the Alan Turing InstituteThink tanksCentre for British Progress2026
19. Alan Turing Institute: UKRI review finds institute underperformingNewsResultSense2026-04
20. Introducing OpenAI LondonIndustryOpenAI2026-04
21. Artificial Intelligence (Regulation) Bill [HL] — Bill 3942GovernmentUK Parliament2026
22. Joint statement on frontier AI models and cyber resilienceGovernmentBank of England / FCA / HM Treasury2026-05-15
23. Government abolishes DSIT as AI gains a seat at the Cabinet tableNewsTHINK Digital Partners2026-07-21
24. Machinery of Government Changes (HLWS298)GovernmentUK Parliament — Written Statements2026-07-21
25. Secretary of State for Business, Innovation, Science and TradeGovernmentGOV.UK2026-07
26. UK Sovereign AI Fund & Bezos Backs CuspAI in £335M Series B FundingNewsdigit.fyi2026-07-20
27. Nobel laureate John Jumper is leaving DeepMind for rival AnthropicNewsTechCrunch2026-06
28. Frontier AI cyber security guidance & risksGovernmentNational Cyber Security Centre (NCSC)2026-07
29. Why did OpenAI's and Anthropic's AI models hack other companies?NewsNPR2026-08
30. Reports: UK could consider regulation amidst growing 'rogue AI' concernsNewsIT Pro2026-08-04
31. Incident Report: unsanctioned agent behaviour during cyber testingGovernmentUK AI Security Institute2026-08-04
32. UK cyber bill targets AI users, not the vendors building itNewsThe Register2026-09-02