GB — Country Profile

United Kingdom

301TOTAL
220OFFICIAL SOURCES
33TOPIC AREAS
Law / Act38
Policy / Guidance15
National Strategy25
Standard / Framework15
International Agreement12
Working Paper25
Court Case51
News / Press17
Other103
AdvertisingAgentic AiAi SafetyAntitrust & CompetitionChips & Data CentresComputeCopyright & Intellectual PropertyCopyright & IpCybersecurityData Privacy & ProtectionDeepfakesDefense & National SecurityEducationEnergy & EnvironmentFinanceGenerative AIHealth & Life SciencesHuman Rights & EthicsInfrastructure & HardwareJudicial & Law EnforcementLabor & WorkforceLiability & AccountabilityMedia & EntertainmentNational StrategyOnline Safety & Child ProtectionPublic Sector & GovernanceResearchSafe & Responsible AiSandboxTelecommunicationsTrade & InvestmentTraining & WorkforceTransport
Policy / Guidance✓ Official

AI Safety Institute: overview and update — Department for Science, Innovation and Technology

This policy paper announced the establishment of the UK AI Safety Institute (AISI) as the world's first government body dedicated to AI safety research and evaluation, launched at the AI Safety Summit at Bletchley Park. The AISI is tasked with conducting safety evaluations of frontier AI models, building the scientific and empirical understanding of AI safety risks, and sharing findings with the international community. It operates within the Department for Science, Innovation and Technology (DSIT) and has subsequently conducted evaluations in partnership with the US AI Safety Institute.

29 November 2023National StrategyCybersecurityHuman Rights & Ethics
↗ Link available
Policy / Guidance✓ Official

AI Regulation: a pro-innovation approach — White Paper (CP 815)

This white paper set out the UK government's initial framework for AI regulation, deliberately rejecting a single binding AI law in favour of a principles-based, sector-led approach administered by existing regulators such as the FCA, CMA, and ICO. It identified five cross-sector principles — safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; contestability and redress — to be applied contextually by sectoral regulators rather than through new primary legislation. A consultation ran from March to June 2023, followed by an updated response paper in February 2024.

29 March 2023National StrategyLiability & AccountabilityHuman Rights & Ethics
↗ Link available
Policy / Guidance✓ Official

A pro-innovation approach to AI regulation: government response (CP 1019)

This document set out the UK government's response to the March 2023 white paper consultation, confirming the decision to embed the five AI principles in existing regulatory frameworks rather than create a standalone AI Act. It announced additional measures including mandatory transparency requirements for AI use in public services, a central risk function, and new regulatory sandboxes, while retaining flexibility to introduce binding obligations for frontier models if voluntary commitments proved insufficient. The response also confirmed the UK's international engagement strategy through the AI Safety Institute and bilateral agreements with allied nations.

06 February 2024National StrategyLiability & AccountabilityCybersecurity
↗ Link available
International Agreement✓ Official

Seoul Declaration for Safe, Innovative and Inclusive AI — Ministerial Statement, AI Seoul Summit 2024

The Seoul AI Safety Summit produced two key outcomes: the Seoul Ministerial Declaration endorsed by 27 nations and the Seoul Statement of Intent on AI Safety, in which major frontier AI developers made voluntary safety commitments. The Ministerial Declaration built on Bletchley by extending commitments to AI safety evaluations, international interoperability of safety standards, and inclusive development that benefits all nations including the Global South. It also launched the international network of AI safety institutes and confirmed plans for a further summit in France in early 2025.

21 May 2024National StrategyCybersecurityHuman Rights & Ethics
↗ Link availableFull text
Working Paper✓ Official

Artificial Intelligence (Regulation) Bill — House of Lords Private Member's Bill [HL Bill 37]

Lord Holmes of Richmond introduced this Private Member's Bill to establish a statutory framework for AI regulation in the UK, including a mandatory registration scheme for high-risk AI systems, a duty of transparency on AI developers, and the creation of a dedicated AI Authority. While the bill did not progress to become law during its session, it represents the most developed attempt to introduce primary AI legislation in the UK Parliament and reflects growing pressure on the government to move beyond its voluntary, principles-based approach. It is distinct from the government's preferred sector-led model and serves as a legislative benchmark for AI governance advocates.

22 October 2024National StrategyLiability & AccountabilityHuman Rights & Ethics
↗ Link available
National Strategy✓ Official

AI Action Plan — Department for Science, Innovation and Technology

The UK government's AI Opportunities Action Plan, published in January 2025, sets out the strategy for making the UK a global AI superpower, focusing on compute infrastructure, data access, talent retention, and adoption across public services. It accepted all 50 recommendations from the independent review led by Matt Clifford, including creation of AI Growth Zones, a National Data Library, and expanded public sector AI deployment. The plan signals a significant shift toward prioritising AI-driven economic growth alongside safety, and is accompanied by a package of investment announcements and international partnerships.

22 January 2025National StrategyLabor & WorkforceEducation
↗ Link available
Court Case✓ Official

Kamal v Tax Policy Associates

Lawyer appeared before the High Court. Fabricated: Case Law | Claimant deployed AI-generated fabricated case citations in correspondence/pleadings, prompting extra cross-checking; court described the deployment as unacceptable and relied on it as part of poor litigation conduct. Outcome: Partial strike-out, summary judgment for the defendants.

Court: High CourtParty: Lawyer
No
11 March 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Re A, B, C, D (Extension of assessment; Use of AI: hallucinations)

Pro Se Litigant used Unidentified in proceedings before the Family Court Bournemouth. Fabricated: Case Law | Skeleton argument contained citations that did not relate to the cases named; advocates agreed several citations/propositions were erroneous and the court identified these as AI-generated fabricated case law.

Court: Family Court BournemouthParty: Pro Se LitigantTool: Unidentified
No
5 March 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Brightwaters Energy Limited v Eroton Exploration and Production Company Limited

Lawyer used Unidentified in proceedings before the High Court. Fabricated: Case Law | Eroton's counsel relied on a supposed 2014 HHJ Pelling KC authority which did not exist; the court held the text was invented by AI and counsel promptly acknowledged the error.

Court: High CourtParty: LawyerTool: Unidentified
No
17 February 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

PSAHSC v. Tchampet

Lawyer used Microsoft Co-Pilot in proceedings before the High Court. Misrepresented: Case Law | Written submissions relied on Ghosh v GMC [2001] UKPC 29 as authority for a proposition it did not support; court found the AI had produced an unsupported citation/authority. Outcome: Warning.

Court: High CourtParty: LawyerTool: Microsoft Co-Pilot
No
30 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Folarin v. The Immigration Services Commissionner

Pro Se Litigant used ChatGPT in proceedings before the First-tier Tribunal. Fabricated: Case Law | Tribunal was unable to locate several case citations the appellant relied on; the citations were generated by ChatGPT and could not be found in any database, suggesting AI-fabricated cases.

Court: First-tier TribunalParty: Pro Se LitigantTool: ChatGPT
No
29 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Your Home Partners v. Kellichan

Pro Se Litigant used Unidentified in proceedings before the Sheriffdom of Tayside. Fabricated: Legal Norm | Claimant lodged extract purporting to be 'Small Claims (Scotland) Rules – Section 41a' which does not exist; court determined the rule was fabricated. Outcome: Warning.

Court: Sheriffdom of TaysideParty: Pro Se LitigantTool: Unidentified
No
19 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Robert Huish v HRMC

Pro Se Litigant used Unidentified in proceedings before the First-tier Tribunal. Fabricated: Case Law | Appellant relied on an alleged 2021 case named 'Collins'; Tribunal and HMRC checked and found no such case after appellant admitted using AI assistance.

Court: First-tier TribunalParty: Pro Se LitigantTool: Unidentified
No
16 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Ms F Green v Imprint Creative Print Solutions Limited

Pro Se Litigant used Unidentified in proceedings before the Employment Tribunal (Leeds). Fabricated: Case Law | Claimant submitted case law authorities in an email that could not be identified; claimant admitted they were AI-generated and the Tribunal found the authorities do not exist and should be disregarded.

Court: Employment Tribunal (Leeds)Party: Pro Se LitigantTool: Unidentified
No
12 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Elden v HMRC

Lawyer used Unidentified in proceedings before the First-tier Tribunal. Misrepresented: Case Law | Skeleton summarized Atlantic Electronics Ltd v HMRC [2012] UKUT 45 (TCC) as if it were a strike-out authority; Tribunal read the full 9-page decision and found the summary inaccurate and irrelevant (case concerned MTIC VAT and admission of additional evidence). Outcome: Order to include statements of truth about AI use.

Court: First-tier TribunalParty: LawyerTool: Unidentified
No
8 January 2026Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

M Peiu v Hywel Dda University Local Health Board

Pro Se Litigant used ChatGPT in proceedings before the Employment Tribunal (Swansea). Fabricated: Case Law | Claimant relied on multiple case citations produced via ChatGPT in her written submissions but could not provide citations or copies; Tribunal and counsel could not locate them and did not rely on those authorities.

Court: Employment Tribunal (Swansea)Party: Pro Se LitigantTool: ChatGPT
No
30 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Mr T De Carvalho Ferreira v Magic Life Limited & Ors

Lawyer used Unidentified in proceedings before the Employment Tribunals (Reading). Fabricated: Case Law | Claimant's skeleton cited 'Adams v Wincanton Group Ltd [2013] IRLR 725 (EAT)'; the judge and opposing counsel were unable to locate the case; counsel acknowledged using AI and was warned to verify AI outputs. Outcome: Warning.

Court: Employment Tribunals (Reading)Party: LawyerTool: Unidentified
No
12 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Mr J Harrison v Mr D May t/a Leeds Gymnastics Academy

Pro Se Litigant used ChatGPT in proceedings before the Employment Tribunals (Leeds). Fabricated: Case Law | Respondent's written submissions contained numerous non-existent case law references; respondent admitted he used ChatGPT to generate the submissions and did not verify citations. Tribunal found at least half of the references were non-existent. Outcome: Tribunal awarded claimant a preparation time order of £2,178; payment stayed pending outcome of EAT appeal..

Court: Employment Tribunals (Leeds)Party: Pro Se LitigantTool: ChatGPT
No$ 2178 GBP
10 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

D (A Child) (Recusal)

Pro Se Litigant used Unidentified in proceedings before the Court of Appeal. Misrepresented: Case Law | Mother cited Re W [2010] UKSC 12 for the proposition that 'findings reached through a procedurally compromised process cannot stand', an attribution the court found to be incorrect. Outcome: Warning.

Court: Court of AppealParty: Pro Se LitigantTool: Unidentified
No
9 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Onyinye Udokporo v Enrich International

Pro Se Litigant used Unidentified in proceedings before the IPO. Fabricated: Case Law | Appellant cited 'Combit Software GmBH v Commit Business Solutions [2014] EWHC 3605' claiming judicial comment on minor consonant changes; court found this citation fabricated and the appellant later supplied a different EU case (C-223/15).

Court: IPOParty: Pro Se LitigantTool: Unidentified
No
5 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Taiwo v Homelets of Bath Limited

Pro Se Litigant appeared before the High Court. Fabricated: Case Law | A purported authority cited in the 27 March 2025 skeleton (requested by the Court) could not be produced; the Court found Irani v Duchy Farm Kennels [2020] EWCA Civ 405 to be a bogus/non-existent case, likely generated by AI. Outcome: Adverse Costs Order.

Court: High CourtParty: Pro Se Litigant
No$ 1
3 December 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Oxford Hotel Investments Ltd v Great Yarmouth Borough Council

Pro Se Litigant used Unidentified in proceedings before the Upper Tribunal (Lands Chamber). False Quotes: Case Law | Appellant cited [18] of Barker v Shokar as holding a microwave satisfied the statutory definition; the Tribunal found the passage did not say that and that the misleading citation was produced by AI.

Court: Upper Tribunal (Lands Chamber)Party: Pro Se LitigantTool: Unidentified
No
24 November 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Mark Jennings v NatWest Group PLC

Pro Se Litigant used ChatGPT in proceedings before the Sheriff Appeal Court. Fabricated: Case Law | Court observed that at least three cases cited in the appellant's ChatGPT-generated submissions appear to be non-existent; the court treated those citations as unreliable.

Court: Sheriff Appeal CourtParty: Pro Se LitigantTool: ChatGPT
No
21 November 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

133 Blackstock Road v Assethold Limited

Lawyer appeared before the First-tier Tribunal (Property Chamber). Fabricated: Case Law | Respondent relied on Gala Unity Ltd v Ariadne Road RTM Co Ltd [2012] UKUT 366 (LC); Tribunal could not find this authority and directed the Respondent to produce it; none was provided and Tribunal concluded it apparently does not exist. Outcome: Show Cause.

Court: First-tier Tribunal (Property Chamber)Party: Lawyer
No$ 1
17 November 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

UK and R v Secretary of State for the Home Department

Lawyer used ChatGPT in proceedings before the Upper Tribunal. Fabricated: Case Law | A non-existent case 'Horleston' was included in the grounds; the Tribunal found no reported case and demonstrated that Google AI can fabricate bench details—solicitor concluded it was probably AI-generated. Outcome: Bar Referral.

Court: Upper TribunalParty: LawyerTool: ChatGPT
Yes
17 November 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Choksi v IPS

Lawyer used Google AI Overview in proceedings before the High Court. Fabricated: Case Law | Farnell 2 cited numerous authorities with wrong citations or names and included some cases that do not exist; court found several citations irrelevant and unsupported by the authorities cited.

Court: High CourtParty: LawyerTool: Google AI Overview
No
4 November 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

OscarTech UK Ltd v ORTHOFIX S.r.l.

Pro Se Litigant used Unidentified in proceedings before the UK IPO. Fabricated: Case Law | Applicant's Skeleton Argument cited a decision as 'IONTECH v TECNION (BL O/375/10)'; Opponent flagged the reference as inconsistent and the Applicant withdrew reliance after admitting use of an AI tool. Outcome: Warning.

Court: UK IPOParty: Pro Se LitigantTool: Unidentified
No
30 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Victoria Place Flats RTM Company Ltd & ors v Assethold Limited

Lawyer appeared before the First-tier Tribunal. Fabricated: Case Law | M365 Copilot returned fabricated Court of Appeal citations for Qdime (non-existent EWCA citations) which Mr Gurvits' submissions echoed; tribunal could not locate those citations on bailii and found them to be AI hallucinations.

Court: First-tier TribunalParty: Lawyer
No
23 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Re Sriram (aka Roy)

Pro Se Litigant appeared before the High Court. Fabricated: Case Law | Litigant relied on a non-existent case citation which the court identified as invented; judge found the citation does not exist and likely arose from internet research error. Outcome: Warning.

Court: High CourtParty: Pro Se Litigant
No
22 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Chi Keung Lee & others v Blackpool B&B Limited

Pro Se Litigant used Unidentified in proceedings before the First-tier Tribunal. Fabricated: Case Law | Respondent submitted a purported Upper Tribunal decision 'Khan v Mehmood' as legal authority; Tribunal found no such decision exists and concluded it was fabricated by AI. Outcome: Monetary Sanction.

Court: First-tier TribunalParty: Pro Se LitigantTool: Unidentified
No$ 227 GBP
17 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Gloriose Ndaryiyumvire v Birmingham City University

Lawyer used LEAP legal software in proceedings before the County Court (Birmingham). Fabricated: Case Law | Application cited 'Qureshi v Qureshi [1998] 1 WLR 174' as authority for amendments; court found no such authority exists and the citation was fictitious (the cited neutral citation corresponds to an unrelated case). Outcome: Wasted costs order; Bar Referral.

Court: County Court (Birmingham)Party: LawyerTool: LEAP legal software
No
14 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Mr M Peters v Driver and Vehicle Standards Agency

Pro Se Litigant used Unidentified in proceedings before the Employment Tribunals (Cambridge). Fabricated: Case Law | AI-generated report for 'Nally v Freshfield Care ET/2401774/14' could not be produced on request and could not be verified; tribunal concluded the summary was unreliable and possibly fabricated.

Court: Employment Tribunals (Cambridge)Party: Pro Se LitigantTool: Unidentified
No
9 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

AK v Secretary of State for the Home Department

Lawyer used ChatGPT in proceedings before the Upper Tribunal. Fabricated: Case Law | Grounds of appeal cited 'R (Goudey) v SSHD [2012] EWHC 567 (Admin)'; Judge found the authority does not exist and treated it as another example of AI-generated false citation. Outcome: Show Cause Order.

Court: Upper TribunalParty: LawyerTool: ChatGPT
No
6 October 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

ANPV & SAPV v Secretary of State for the Home Department

Lawyer used Microsoft Copilot in proceedings before the Upper Tribunal. Fabricated: Case Law | Grounds cited a Honduras country guidance decision that does not exist; judge found the MN Honduras CG [2016] authority fabricated. Outcome: Show Cause Order.

Court: Upper TribunalParty: LawyerTool: Microsoft Copilot
No
29 September 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Kuzniar v General Dental Council

Pro Se Litigant used ChatGPT in proceedings before the Employment Tribunal. Fabricated: Case Law | Claimant's skeleton arguments and authority bundles contained 15 authorities identified by Respondent as non-existent (mixes of real names and non-existent citations); Respondent could not locate them. Outcome: Tribunal declined to award costs.

Court: Employment TribunalParty: Pro Se LitigantTool: ChatGPT
No
15 August 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

MS (Bangladesh)

Lawyer used ChatGPT in proceedings before the Immigration and Asylum Chamber. Fabricated: Case Law | Grounds of appeal cited a non-existent Court of Appeal authority 'Y (China)' as though it dealt with undue weight on delay; Tribunal found the case does not exist and was generated by ChatGPT. Outcome: Referred to the Bar Standards Board for investigation.

Court: Immigration and Asylum ChamberParty: LawyerTool: ChatGPT
Yes
12 August 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Holloway v Beckles

Pro Se Litigant used Unidentified in proceedings before the First-tier Tribunal. Fabricated: Case Law | Tribunal noted the fake cases were cited alongside some genuine authorities, but the fabricated citations amounted to unreasonable conduct for costs purposes. Outcome: Costs Order.

Court: First-tier TribunalParty: Pro Se LitigantTool: Unidentified
No$ 750 GBP
12 August 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

The Father v The Mother

Pro Se Litigant appeared before the High Court. Fabricated: Case Law | Father's written application cited multiple authorities relating to ASD which HHJ Bailey concluded were not genuine and appeared to have been generated by AI; judge noted the Father 'relied upon faked cases' without checking veracity. Outcome: Costs awarded against the Father (partly for relying on faked cases)..

Court: High CourtParty: Pro Se Litigant
No$ 5900 GBP
30 July 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

Chanda v Royal Mail Group Ltd

Lawyer appeared before the Employment Tribunal. Misrepresented: Case Law | Tribunal could not find the case or the propositions relied on; claimant accepted the propositions cited could not be found.

Court: Employment TribunalParty: Lawyer
No
28 July 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available
Court Case✓ Official

HMRC v. Gunnarson

Pro Se Litigant used Unidentified in proceedings before the Upper Tribunal (Tax and Chancery Chamber). Fabricated: Case Law | Respondent’s skeleton cited a non-existent FTT authority said to show HMRC’s SEISS guidance was unclear; UT records HMRC checked and the case does not exist. Outcome: Warning.

Court: Upper Tribunal (Tax and Chancery Chamber)Party: Pro Se LitigantTool: Unidentified
No
23 July 2025Judicial & Law EnforcementGenerative AILiability & Accountability
↗ Link available