At a glance
- As of November 3, 2026, the UK has no general AI statute. The government’s 2023 white paper set five cross-sector principles on a non-statutory basis, for existing regulators to apply in their own sectors.
- Binding rules reach AI through existing law. Since February 5, 2026, new UK GDPR Articles 22A to 22D govern significant decisions based solely on automated processing. The Online Safety Act 2023 applies to regulated user-to-user and search services, and the Equality Act 2010 applies in England, Wales and Scotland. In England and Wales, making or supplying a tool for creating purported intimate images has been a criminal offence since June 29, 2026.
- No government AI bill had been introduced as of November 3, 2026. The 2024 King’s Speech said the government would seek legislation to place requirements on those working to develop the most powerful AI models. The May 13, 2026 King’s Speech briefing notes list no AI bill.
- The Artificial Superintelligence Bill (bill 4288), a Ten Minute Rule private member’s bill, had its Commons first reading on September 8, 2026. As of November 12, 2026, its second reading, first listed for November 13, 2026, was listed for November 27, 2026, a date that may be provisional, and no bill text was published. It is not law.
Does the UK have an AI law?
No general one. As of November 3, 2026, the UK has no statute that regulates AI systems as such. AI is regulated through existing law, chiefly data protection, online safety, equality, criminal, competition, financial services and medical device law. Each sector’s regulator applies that law, guided by five principles the government’s 2023 white paper issued on a non-statutory basis.
The rules attach to the activity, not to a risk tier: a solely automated decision about a person, content on an online service, a bank’s model, a medical device. The UK GDPR can reach a company outside the UK. Under Article 3(2), it applies to a controller or processor not established in the UK whose processing of personal data of people in the UK relates to offering them goods or services or to monitoring their behaviour there.
What did the white paper and the 2024 response set up?
The Department for Science, Innovation and Technology (DSIT) published the white paper, A pro-innovation approach to AI regulation, on March 29, 2023. It set five principles: safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress. The white paper said the principles “will be issued on a non-statutory basis and implemented by existing regulators,” and that the non-statutory approach “will not create new rights or new routes to redress at this stage.”
The government response of February 6, 2024 kept that approach. It said a non-statutory approach “currently offers critical adaptability” and that the government would keep a statutory duty on regulators under review. It set out the case for “targeted, binding requirements on developers of highly capable general-purpose AI models in the future,” announced £10 million for regulators’ AI capabilities, and asked key regulators to publish their strategic approach to AI by April 30, 2024.
On May 1, 2024, DSIT published updates from 13 regulators, among them the Bank of England, CMA, Equality and Human Rights Commission, FCA, ICO, MHRA and Ofcom.
Two later steps changed emphasis, not law. The AI Opportunities Action Plan, an independent report led by Matt Clifford with 50 recommendations, was published on January 13, 2025. On February 14, 2025, the government renamed the AI Safety Institute the AI Security Institute; the announcement says it will focus on serious AI risks with security implications and “will not focus on bias or freedom of speech.”
Which UK rules apply to AI in 2026?
Read the “Binding or voluntary” column first. The table mixes statutes in force, a regulation-making power with no regulations made under it yet, supervisory expectations, voluntary programmes and a policy report.
| Area | Law or guidance | Regulator | Binding or voluntary | Status or date |
|---|---|---|---|---|
| Cross-sector principles | AI regulation white paper and 2024 government response | Existing regulators, coordinated by DSIT | Non-statutory | March 29, 2023; response February 6, 2024 |
| Automated decisions | UK GDPR Articles 22A to 22D, inserted by Data (Use and Access) Act 2025, s. 80 | Information Commission | Binding | In force February 5, 2026 |
| High-risk processing | UK GDPR Article 35 data protection impact assessment | Information Commission | Binding | In force |
| AI and automated decisions code | Data Protection Act 2018 (Code of Practice on Artificial Intelligence and Automated Decision-Making) Regulations 2026, S.I. 2026/425 | Information Commission | Duty on the regulator to prepare a code; the code will be guidance | Regulations in force May 12, 2026 |
| Online content | Online Safety Act 2023, including content users generate, upload or share by means of software or an automated tool (s. 55(4)) | Ofcom | Binding on regulated services | In force; 48-hour takedown duty for reported intimate image content since June 29, 2026 (s. 10(3A)) |
| AI services | Online Safety Act 2023, s. 216A power on illegal AI-generated content and the use of AI services for priority offences | Secretary of State (regulations); Ofcom | Power only; no regulations found | Inserted April 29, 2026; progress report due by December 31, 2026 |
| Intimate image tools | Sexual Offences Act 2003, s. 66E (creating a purported intimate image of an adult without consent) and s. 66I (making or supplying a purported intimate image generator) | Police, prosecutors and courts | Binding (criminal offences), England and Wales | s. 66E in force February 6, 2026; s. 66I in force June 29, 2026 |
| Discrimination | Equality Act 2010, ss. 13, 19, 29, 39 | Equality and Human Rights Commission; courts and tribunals | Binding | In force in England, Wales and Scotland |
| Financial conduct | FCA Principles, Consumer Duty, SM&CR; AI Live Testing | FCA | Rules binding; AI Live Testing voluntary | Second cohort announced April 21, 2026 |
| Bank model risk | PRA Supervisory Statement SS1/23 | Prudential Regulation Authority | Supervisory expectations | Effective May 17, 2024; current version April 23, 2026 |
| Digital markets | Digital Markets, Competition and Consumers Act 2024; Google search designation and conduct requirement | CMA | Binding on the designated firm | Designation October 10, 2025; conduct requirement June 3, 2026 |
| Medical devices | Medical device regulation of software and AI; AI Airlock sandbox | MHRA | Device rules binding; Airlock reports are not formal guidance | Airlock phase 3 launched October 6, 2026 |
| Copyright and AI training | Copyright, Designs and Patents Act 1988; report under Data (Use and Access) Act 2025, s. 136 | DSIT, DCMS and Intellectual Property Office | Report is policy; it proposes no change to copyright law for now | Report published March 18, 2026 |
Is there a UK AI bill?
No government bill. The July 17, 2024 King’s Speech said the government “will seek to establish the appropriate legislation to place requirements on those working to develop the most powerful artificial intelligence models.” A title search of bills.parliament.uk on November 3, 2026 found no government bill on AI in the 2024–26 or 2026–27 sessions.
The background briefing notes for the King’s Speech of May 13, 2026 list no AI bill. They describe a Regulating for Growth Bill that would create cross-economy “sandboxing powers” to let existing rules be temporarily relaxed, under strict controls, to test new products and technologies, including “exploring cross-cutting AI sandboxes.” No bill with that title appeared on bills.parliament.uk on November 3, 2026.
Title searches of bills.parliament.uk for “artificial intelligence” and “superintelligence” found these private member’s bills in the 2024–26 and 2026–27 sessions. None is law.
| Bill | Type and sponsor | Furthest stage | Status |
|---|---|---|---|
| Artificial Superintelligence Bill (bill 4288), 2026–27 | Private member’s bill (Ten Minute Rule); Alex Sobel MP (Labour), House of Commons | Commons first reading, September 8, 2026 | Second reading listed for November 27, 2026, moved from November 13 (future dates may be provisional); no bill text published as of November 12, 2026 |
| Artificial Intelligence (Consent and Misuse) Bill, 2026–27 | Private member’s bill (presentation bill); Dawn Butler MP (Labour), House of Commons | Commons first reading, October 20, 2026 | Second reading listed for November 27, 2026 (future dates may be provisional); no bill text published |
| Artificial Intelligence (Regulation) Bill [HL], 2024–26 | Private member’s bill starting in the House of Lords; Lord Holmes of Richmond (Conservative) | Lords first reading, March 4, 2025 | Fell at the end of the session; no 2026–27 version found |
The long title of bill 4288 says it would “prohibit the development, deployment and operation of artificial superintelligence systems” and establish monitoring and control powers. With no text published, its definitions and duties cannot be assessed. SuperintelligenceDefinition.com (related publication) compares it with the US ban bill in its superintelligence ban guide.
The long title of the Consent and Misuse Bill says it would make provision about a requirement for consent to generate a likeness of a person using, or with the assistance of, AI; about civil action where that requirement is not met; about offences relating to creating or disseminating such likenesses; and about duties on providers of AI technology in relation to misuse of that technology to generate likenesses.
Two other private member’s bills reach automated systems without naming AI in their titles. The Public Authority Algorithmic and Automated Decision-Making Systems Bill [HL] (Lord Clement-Jones, 2024–26), which would have required public authorities to complete impact assessments of automated and algorithmic decision-making systems, completed its Lords stages on February 7, 2025 and fell when the session ended on April 29, 2026. The Automated Online Software (Access and Transparency) Bill (Damian Hinds, 2026–27, a ballot bill) would make provision about registering operators of specified automated online software and about their duties, including on accessing online material published by others; as of November 3, 2026, its second reading was listed for January 15, 2027.
What does UK data protection law require for automated decisions?
Two tests in Article 22A decide whether the new rules reach a decision: meaningful human involvement in taking it, and significance for the data subject. Section 80 of the Data (Use and Access) Act 2025, which received Royal Assent on June 19, 2025, substituted new Articles 22A to 22D for Article 22 of the UK GDPR. It came into force on February 5, 2026 under S.I. 2026/82, regulation 2(j). Regulation 5 keeps the old rules for decisions taken before that date to which Article 22(3) of the UK GDPR, or section 14 or 50(2) of the Data Protection Act 2018, applied.
| Article | What it says |
|---|---|
| 22A (scope) | A decision is based solely on automated processing “if there is no meaningful human involvement in the taking of the decision.” It is significant if it produces a legal effect or has a similarly significant effect for the data subject. In judging meaningful human involvement, a person must consider, among other things, the extent to which the decision is reached by means of profiling. |
| 22B(1) to (3) (special category data) | A significant decision based entirely or partly on special category data may not be taken based solely on automated processing unless (1) the decision is based entirely on processing to which the data subject has given explicit consent, or (2) the decision is necessary for entering into or performing a contract between the data subject and a controller, or required or authorised by law, and point (g) of Article 9(2) applies. |
| 22B(4) (recognised legitimate interests) | A significant decision may not be taken based solely on automated processing if the processing by or on behalf of the decision-maker for the decision relies entirely or partly on Article 6(1)(ea). |
| 22C (safeguards) | For a significant decision based entirely or partly on personal data and based solely on automated processing, the controller must ensure safeguards for the data subject’s rights, freedoms and legitimate interests that comply with 22C(2) and any regulations under Article 22D(3). They must include measures that give the data subject information about such decisions and let the data subject make representations, obtain human intervention on the part of the controller, and contest the decisions. |
| 22D (regulations) | The Secretary of State may provide by regulations when there is, or is not, meaningful human involvement, which decisions do or do not have a similarly significant effect, and further or supplementary safeguard requirements. Regulations may not amend Article 22C. |
Infringing Article 22B or 22C is in the top fine tier: up to £17,500,000 or, for an undertaking, up to 4% of total worldwide annual turnover of the preceding financial year, whichever is higher (UK GDPR Article 83(5)(ba)).
Article 35 still requires a data protection impact assessment before processing “in particular using new technologies” that is likely to result in a high risk to the rights and freedoms of natural persons, and in particular for systematic and extensive evaluation of personal aspects based on automated processing, including profiling, on which decisions with legal or similarly significant effects are based.
The regulator changed its form on September 30, 2026: the office of Information Commissioner was abolished and its functions transferred to the Information Commission (Data (Use and Access) Act 2025, ss. 118 and 119). A reference to the Information Commissioner in an enactment or other document is treated, so far as appropriate, as a reference to the Commission. The government’s 2024 response cited the ICO’s updated guidance on data protection and AI.
A statutory code on AI is in preparation. S.I. 2026/425, in force since May 12, 2026, requires the regulator to prepare a code of practice on good practice in processing personal data when developing and using AI and in automated decision-making, including guidance on children’s data; the duty passed to the Information Commission on September 30, 2026. The government’s explanatory memorandum describes the code as “authoritative, non-binding guidance” and says it will be subject to parliamentary scrutiny before it is issued.
Does the Online Safety Act cover generative AI?
In part. The Act regulates user-to-user services and search services (section 3). For user-to-user services, content generated, uploaded or shared “by means of software or an automated tool applied by the user” counts as user-generated content, and a bot is treated as a user if its functions include interacting with user-generated content and it is not controlled by or on behalf of the provider (section 55(4)). Ofcom can fine a provider up to the greater of £18 million and 10% of qualifying worldwide revenue (Schedule 13, paragraph 4).
Since June 29, 2026, a regulated user-to-user service must operate proportionate systems and processes designed to take down content reported to it as intimate image content, and content it identifies as the same or substantially the same, as soon as reasonably practicable and no later than 48 hours after receiving the report, unless the provider considers that the content is not intimate image content or that the reporter is not the person shown or someone acting for that person (section 10(3A) and (3B), inserted by section 100 of the Crime and Policing Act 2026). Regulated search services have a parallel duty (section 27(3A)). Intimate image content is content that amounts to an offence of sharing an intimate photograph or film without consent under section 66B(1) to (3) of the Sexual Offences Act 2003, which covers an image made or altered by computer graphics that appears to be a photograph or film (sections 66A(5) and 66D(4)), so AI-generated images can qualify. The deepfake laws guide sets this beside the US TAKE IT DOWN Act.
Not every chatbot is in scope, by the government’s own account. On February 15, 2026, it said it was tabling an amendment to the Crime and Policing Bill to allow the government to require chatbots not currently in scope of the Online Safety Act to protect their users from illegal content. The Crime and Policing Act 2026, with Royal Assent on April 29, 2026, inserted section 216A into the Online Safety Act. It lets the Secretary of State, by regulations, amend the Act to minimise or mitigate risks of harm to individuals in the UK from illegal AI-generated content and from the use of AI services to commit or facilitate priority offences, including by imposing illegal content duties on providers of AI services. Section 249 of the 2026 Act requires a progress report to Parliament by December 31, 2026 unless draft regulations are laid first. No regulations under section 216A appeared on legislation.gov.uk on November 3, 2026.
The government’s July 2026 response to its Growing up in the online world consultation says it will prevent under-18s from accessing AI chatbot services that primarily offer sexualised content. That is a policy commitment, not law.
The criminal law already reaches some AI tools. In England and Wales, section 66I of the Sexual Offences Act 2003, inserted by section 99 of the Crime and Policing Act 2026 and in force since June 29, 2026, makes it an offence to make or adapt a thing, or to supply or offer to supply it, for use as a generator of purported intimate images; a “thing” includes a program and a service. It is a defence to prove that all reasonable steps were taken to prevent the thing being used to create purported intimate images of a person without that person’s consent. The maximum penalty on indictment is 3 years’ imprisonment, and where a company commits the offence with the consent or connivance of a director, manager or similar officer, that person commits it too. Creating a purported intimate image of an adult who does not consent, without reasonable belief in consent, has been an offence since February 6, 2026 (section 66E).
What do financial, competition and health regulators expect?
These regulators mostly apply rules they already had. The one new binding measure in this table, the CMA’s conduct requirement, binds Google alone.
| Regulator | Instrument | What it says about AI | Status |
|---|---|---|---|
| FCA | AI Update (April 2024) | The FCA calls itself “a technology-agnostic, principles-based and outcomes-focused regulator.” Its framework “does not specifically address the transparency or explainability of AI systems,” and it points to the Principles for Businesses, the Consumer Duty and the Senior Managers and Certification Regime as relevant to firms’ use of AI. Its current AI page says it does not plan to introduce extra regulations for AI and will rely on existing frameworks. | Existing rules bind |
| FCA | AI Live Testing | Participating firms receive tailored support from the FCA and its technical partner Advai “to develop, assess and deploy safe and responsible AI.” | Voluntary; second cohort announced April 21, 2026; testing to end by the end of 2026, evaluation report in Q1 2027 |
| PRA | SS1/23, Model risk management principles for banks | Relevant to UK-incorporated banks, building societies and PRA-designated investment firms with internal model approval to calculate regulatory capital for credit, market or counterparty credit risk. The principles apply to all types of models used to inform business decisions, including vendor models; a sub-principle covers AI risks to the extent that it applies to the use of models more generally. | Supervisory expectations; effective May 17, 2024; current version published and effective April 23, 2026 |
| CMA | AI Foundation Models update paper | Confirms final principles for guiding the foundation model market to positive outcomes for competition and consumer protection. | Published April 11, 2024 |
| CMA | Google general search SMS designation (Digital Markets, Competition and Consumers Act 2024, s. 2) | AI Overviews and AI Mode are in scope; the Gemini assistant is not, a position the CMA will keep under review. The designation is “not a finding of wrongdoing” and introduced no immediate requirements. | October 10, 2025 |
| CMA | Publisher conduct requirement | Among other things, Google must give publishers effective controls over the use of their search content in generative AI and take reasonable steps to ensure that search content is attributed clearly and accurately in general search. | Binding on Google; imposed June 3, 2026 |
| MHRA | Software and AI as a medical device | Many software and AI products are regulated as medical devices or in vitro diagnostic medical devices; the change programme roadmap dates from October 17, 2022. | Device rules bind |
| MHRA | AI Airlock regulatory sandbox | Launched in spring 2024; phase 2 completed in May 2026. The MHRA says the Airlock reports do not constitute formal guidance. | Phase 3 announced October 6, 2026 |
Does the Equality Act 2010 apply to AI decisions?
Yes, without naming AI. Sections 13, 19, 29 and 39 extend to England, Wales and Scotland. Direct discrimination is treating a person less favourably because of a protected characteristic (section 13). Indirect discrimination is applying to someone a provision, criterion or practice that puts, or would put, people who share their relevant protected characteristic at a particular disadvantage compared with people who do not, and puts, or would put, that person at that disadvantage, where whoever applies it cannot show it to be a proportionate means of achieving a legitimate aim (section 19). An employer must not discriminate in the arrangements it makes for deciding to whom to offer employment (section 39(1)(a)), and a service-provider must not discriminate as to the terms on which it provides a service (section 29(2)(a)). Neither section exempts treatment or criteria applied by means of software.
Where does UK copyright law stand on AI training?
The government has stepped back from the reform it consulted on. Its March 18, 2026 report concludes that “a broad copyright exception with opt-out is no longer the government’s preferred way forward” and that the government will not introduce reforms to copyright law until it is confident that they will meet its objectives. The report notes that the UK’s existing data mining exception applies to non-commercial research only.
The Copyright and AI consultation ran from December 17, 2024 to February 25, 2025 with four policy options; the government’s preferred option was a broad data mining exception that right holders could opt out of. Sections 135 and 136 of the Data (Use and Access) Act 2025 required an economic impact assessment and a report within nine months of Royal Assent, and the government published both on March 18, 2026.
How does the EU AI Act affect UK firms that sell into the EU?
Leaving the EU did not take UK firms out of the AI Act’s reach. Article 2 covers providers placing AI systems on the EU market or putting them into service there, and providers placing general-purpose AI models on the EU market, wherever they are established. It also covers providers and deployers in third countries where the output of the system is used in the EU. A provider of a general-purpose AI model established outside the EU must appoint an authorised representative in the EU before placing the model on the EU market, a duty that has applied since August 2, 2025, subject to an open-source exemption for models without systemic risk (Article 54). Article 22 sets the same duty for providers of high-risk AI systems established outside the EU, from December 2, 2027 for stand-alone high-risk systems under the Digital Omnibus on AI.
| Point | United Kingdom | European Union |
|---|---|---|
| General AI statute | None | AI Act, Regulation (EU) 2024/1689, as amended by the Digital Omnibus on AI |
| Who applies the rules | Existing sector regulators | National authorities and the AI Office, which supervise and enforce the Act from August 2, 2026 |
| Risk tiers | None in law | Prohibited practices, high-risk systems, systems with transparency duties, general-purpose AI models, and minimal-risk systems |
| Automated decisions | UK GDPR Articles 22A to 22D | GDPR Article 22, plus AI Act duties for high-risk systems |
| Transparency duties for AI-generated content | No AI statute to set them | Article 50, applying from August 2, 2026 |
| Key dates | Articles 22A to 22D in force February 5, 2026 | Stand-alone high-risk duties December 2, 2027; Annex I Section A products August 2, 2028 |
A UK firm that complies with UK law has not thereby complied with the AI Act. Map each EU-facing system to its AI Act role and risk class separately.
What should a company using AI in the UK do now?
- List every automated decision about a person that has a legal or similarly significant effect, and record whether there is meaningful human involvement in taking it
- For each significant decision based solely on automated processing, put in place the Article 22C safeguards set out in the table above
- Check that no such decision relies, entirely or partly, on recognised legitimate interests (Article 6(1)(ea)), and check the Article 22B conditions where special category data is used
- Carry out a data protection impact assessment before AI processing that is likely to result in a high risk to the rights and freedoms of individuals, and name the Information Commission when you next revise privacy notices
- If you run a user-to-user service, check which AI-assisted content and bots count as user-generated content or users under section 55(4) of the Online Safety Act, take down reported intimate image content, including AI-generated images, as soon as reasonably practicable and within 48 hours (section 10(3A); section 27(3A) for search services), and track any section 216A regulations for AI services
- If you make or supply image-generation tools used in England and Wales, check them against section 66I of the Sexual Offences Act 2003 and record the steps you take to prevent non-consensual intimate images
- Test screening, pricing and eligibility models for indirect discrimination and document the legitimate aim and proportionality
- In a bank within SS1/23’s scope, identify and classify AI and vendor models under Principle 1 (model identification and model risk classification); in FCA-regulated firms, map AI use to the Principles for Businesses, the Consumer Duty where it applies, and senior manager responsibilities under SM&CR
- Track the pending private member’s bills (bill 4288 and the Consent and Misuse Bill) and any government AI or sandbox bill on bills.parliament.uk; none is law
- For EU sales, classify each system under the AI Act and appoint an EU authorised representative where Article 22 or Article 54 requires one
How this fits the other guides: the AI compliance guide is the hub for cross-border AI laws. For EU duties, read the EU AI Act guide and the Article 50 transparency guide; compare SS1/23 and the FCA’s approach with the financial services guide, and the UK’s regulator-led model with Canada’s.
This guide is general information as of November 3, 2026, not legal advice. Confirm the current text with the regulator or a solicitor qualified in the relevant part of the UK before acting.
Common questions
Does the UK have an AI law?
No general one. As of November 3, 2026, the UK has no statute that regulates AI systems as such; the 2023 white paper set five principles on a non-statutory basis for existing regulators to apply. Binding rules reach AI through existing law, including the UK GDPR, the Online Safety Act 2023 and the Equality Act 2010.
What are UK GDPR Articles 22A to 22D?
They replaced Article 22 of the UK GDPR from February 5, 2026, under section 80 of the Data (Use and Access) Act 2025, with the old rules kept for certain decisions taken before that date. Where a significant decision based entirely or partly on personal data is taken based solely on automated processing, the controller must ensure safeguards, including measures that give the data subject information about the decision and let the data subject make representations, obtain human intervention on the part of the controller, and contest it; special category data and reliance on recognised legitimate interests carry further restrictions.
What is the UK Artificial Superintelligence Bill?
Bill 4288 is a Ten Minute Rule private member's bill sponsored by Alex Sobel MP. Its long title says it would prohibit the development, deployment and operation of artificial superintelligence systems. It had its Commons first reading on September 8, 2026. As of November 12, 2026, its second reading, first listed for November 13, 2026, was listed for November 27, 2026, a date that may be provisional, and no bill text was published. It is not law.
By industry: all 8 industry pages
Latest: UK AI regulation news, each entry dated and sourced.
Sources
- A pro-innovation approach to AI regulation (white paper), Department for Science, Innovation and Technology and Office for Artificial Intelligence, March 29, 2023.
- A pro-innovation approach to AI regulation: government response, Department for Science, Innovation and Technology, February 6, 2024.
- Regulators’ strategic approaches to AI, Department for Science, Innovation and Technology, May 1, 2024.
- AI Opportunities Action Plan, Department for Science, Innovation and Technology, January 13, 2025.
- Tackling AI security risks to unleash growth and deliver Plan for Change, Department for Science, Innovation and Technology and AI Security Institute, February 14, 2025.
- The King’s Speech 2024: background briefing notes (PDF), Prime Minister’s Office, July 17, 2024.
- The King’s Speech 2026: background briefing notes (PDF), Prime Minister’s Office, May 13, 2026.
- Artificial Superintelligence Bill: stages, UK Parliament, accessed November 12, 2026.
- Artificial Intelligence (Consent and Misuse) Bill: stages, UK Parliament, accessed November 3, 2026.
- Artificial Intelligence (Regulation) Bill [HL], session 2024–26, UK Parliament, accessed November 3, 2026.
- Public Authority Algorithmic and Automated Decision-Making Systems Bill [HL], session 2024–26, UK Parliament, accessed November 3, 2026.
- Automated Online Software (Access and Transparency) Bill, UK Parliament, accessed November 3, 2026.
- Data (Use and Access) Act 2025, section 80, legislation.gov.uk, accessed November 3, 2026.
- The Data (Use and Access) Act 2025 (Commencement No. 6 and Transitional and Saving Provisions) Regulations 2026, S.I. 2026/82, legislation.gov.uk, made January 29, 2026.
- UK GDPR, Article 3, legislation.gov.uk, accessed November 3, 2026.
- UK GDPR, Article 83, legislation.gov.uk, accessed November 3, 2026.
- UK GDPR, Article 35, legislation.gov.uk, accessed November 3, 2026.
- The Data Protection Act 2018 (Code of Practice on Artificial Intelligence and Automated Decision-Making) Regulations 2026, S.I. 2026/425, legislation.gov.uk, made April 16, 2026; see also the explanatory memorandum.
- Data (Use and Access) Act 2025, section 119, legislation.gov.uk, accessed November 3, 2026; see also section 118.
- Online Safety Act 2023, section 55, legislation.gov.uk, accessed November 3, 2026; see also section 3.
- Online Safety Act 2023, Schedule 13, paragraph 4, legislation.gov.uk, accessed November 3, 2026.
- Online Safety Act 2023, section 216A, legislation.gov.uk, accessed November 3, 2026.
- Crime and Policing Act 2026, section 249, legislation.gov.uk, accessed November 3, 2026.
- Crime and Policing Act 2026, section 100, Taking down intimate image content, legislation.gov.uk, accessed November 3, 2026; see also Online Safety Act 2023, section 10 and section 59.
- Sexual Offences Act 2003, section 66I, Making or supplying purported intimate image generators, legislation.gov.uk, accessed November 3, 2026; see also section 66E and S.I. 2026/689.
- PM: “No platform gets a free pass”: Government takes action to keep children safe online, Prime Minister’s Office and others, February 15, 2026.
- Growing up in the online world: government response (July 2026), GOV.UK, updated August 19, 2026.
- Equality Act 2010, section 19, legislation.gov.uk, accessed November 3, 2026; see also section 13, section 29 and section 39.
- AI Update (PDF), Financial Conduct Authority, April 2024; see also AI and the FCA: our approach, Financial Conduct Authority, updated October 2, 2026.
- FCA helps firms to test AI safely, Financial Conduct Authority, December 3, 2025.
- FCA announces second cohort for AI Live Testing, Financial Conduct Authority, April 21, 2026.
- SS1/23: Model risk management principles for banks, Prudential Regulation Authority, Bank of England, May 17, 2023; current version April 23, 2026.
- AI Foundation Models: update paper, Competition and Markets Authority, April 11, 2024.
- Digital Markets, Competition and Consumers Act 2024, section 2, legislation.gov.uk, accessed November 3, 2026.
- CMA confirms Google has strategic market status in search services, Competition and Markets Authority, October 10, 2025.
- Google search publisher conduct requirement, Competition and Markets Authority, June 3, 2026.
- Software and artificial intelligence (AI) as a medical device, Medicines and Healthcare products Regulatory Agency, updated February 3, 2025.
- AI Airlock: a regulatory sandbox for AI enabled Medical Devices, Medicines and Healthcare products Regulatory Agency, updated October 6, 2026.
- Copyright and Artificial Intelligence (consultation), Intellectual Property Office, Department for Science, Innovation and Technology and Department for Culture, Media and Sport, December 17, 2024.
- Data (Use and Access) Act 2025, section 136, legislation.gov.uk, accessed November 3, 2026; see also section 135.
- Report on Copyright and Artificial Intelligence, Department for Science, Innovation and Technology, Department for Culture, Media and Sport and Intellectual Property Office, March 18, 2026.