Glossary

Statutes & legal frameworks

Online Safety Act 2023 (“OSA”)

UK statute imposing duties of care on user-to-user services and search services in respect of illegal content and content harmful to children. Enforced by Ofcom. Drafted before generative AI achieved its present prevalence; FAIR’s position is that it engages only by extension with AI services that themselves speak. Read the Act.

UK GDPR

The retained version of the EU General Data Protection Regulation, supplemented by the Data Protection Act 2018. Regulates the processing of personal data in the United Kingdom. Enforced by the Information Commissioner’s Office.

Article 22 UK GDPR

The right not to be subject to a decision based solely on automated processing which produces legal or similarly significant effects. The basis on which automated decision-making by AI services is regulated in the UK; structurally limited to decisions, and does not engage AI services that speak, advise, or generate.

Data (Use and Access) Act 2025 (“DUAA”)

UK statute amending parts of the UK GDPR including article 22; expanded the scope of automated decision-making rights to certain semi-automated decisions and clarified lawful bases.

Crime and Policing Act 2026

UK statute receiving Royal Assent on 29 April 2026. Sections 38 to 41 made it an offence to possess, create, or distribute AI models fine-tuned on child sexual abuse imagery, or to possess manuals on doing so.

Children’s Wellbeing and Schools Act 2026

UK statute (Royal Assent 29 April 2026) reinforcing safeguarding duties in education and amending Keeping Children Safe in Education guidance.

Charities Act 2011, section 3

The statutory list of charitable purposes in England and Wales. Section 3(1)(b) is “the advancement of education” and section 3(1)(h) is “the advancement of human rights, conflict resolution or reconciliation”. FAIR’s charitable purposes fall under both heads.

EU AI Act (Regulation 2024/1689)

The European Union’s horizontal AI regulation. Article 5 prohibits certain manipulative AI practices; article 50 requires transparency in AI-generated content and AI-mediated interaction. Not directly applicable to UK but a comparator framework.

Article 2 ECHR

The right to life under the European Convention on Human Rights. Engaged in the United Kingdom by the Human Rights Act 1998. Imposes positive obligations on the State to protect life, including in the design of regulatory frameworks.

Communications Decency Act 1996, section 230 (US)

US statute immunising online platforms from liability for content posted by their users. Often invoked as a comparator in AI policy debates; recent US authority indicates that the immunity does not extend to platforms’ own algorithmic editorial speech.

Equality Act 2010

UK statute consolidating equality and anti-discrimination law. Section 19 covers indirect discrimination, relevant to bias in AI-driven decisions.

Leading cases

Garcia v Character Technologies Inc (MD Fla, complaint filed 22 October 2024)

The wrongful-death claim brought by Megan Garcia following the death of her 14-year-old son Sewell Setzer III in February 2024 after a months-long relationship with a Character.AI chatbot persona. Settled by mediation in January 2026. The pleaded facts include the bot’s response “Please do, my sweet king” in the final exchange before his death.

Raine v OpenAI Inc (San Francisco County Superior Court, complaint filed 26 August 2025)

The wrongful-death claim brought by Matthew and Maria Raine following the death of their 16-year-old son Adam in April 2025 after months in which ChatGPT provided him with detailed methods of suicide. Adam Raine’s father gave testimony to the US Senate Judiciary Committee in September 2025.

SCHUFA (Case C-634/21, CJEU 7 December 2023)

The Court of Justice of the European Union ruled that automated credit scoring constituted a decision “based solely on automated processing” for article 22 GDPR purposes, even where a third party formally took the decision relying on the score. Persuasive in UK construction of article 22 UK GDPR.

Moody v NetChoice LLC (603 US 707, US Supreme Court 2024)

US Supreme Court authority on the constitutional status of platforms’ editorial choices. Read by commentators as supporting the view that section 230 immunity does not extend to platforms’ own algorithmic speech.

Regulators & public bodies

Ofcom

The UK’s communications regulator. Lead regulator under the Online Safety Act 2023; publishes guidance on online safety duties and codes of practice.

Information Commissioner’s Office (“ICO”)

The UK’s data protection regulator. Issues codes of practice, statutory guidance, and enforcement notices under the UK GDPR and Data Protection Act 2018. Currently consulting on draft guidance for automated decision-making and profiling, to which FAIR is responding.

Internet Watch Foundation (“IWF”)

The United Kingdom’s sole authorised reporter of online child sexual abuse material. Its 2026 report Harm Without Limits recorded a 26,385 per cent increase in AI-generated child sexual abuse videos in 2025 compared with 2024.

Department for Science, Innovation and Technology (“DSIT”)

UK Government department with policy lead on AI, online safety, and digital regulation. Currently consulting on ‘Growing up in the online world: a national conversation’, to which FAIR is responding.

Charity Commission for England and Wales

The independent regulator of charities in England and Wales. FAIR’s charity registration is pending. The Commission’s guidance Speaking out: guidance on campaigning and political activity by charities (CC9) governs the boundary between charitable advocacy and impermissible political purpose.

Intellectual Property Office (“IPO”)

UK Government office with responsibility for copyright, patents, and trade marks. Conducted review of AI and copyright with report and impact assessment dated 18 March 2026.

NCMEC (US)

The National Center for Missing & Exploited Children, the US-based body to which 18 USC s 2258A obliges electronic communications providers to report apparent child sexual abuse material.

Technical terms

AI chatbot

A software service which produces text or speech responses to user prompts, typically built on a large language model (LLM). General-purpose chatbots, AI companion services, and multimodal generative services are within scope of FAIR’s briefings.

Foundation model

A general-purpose AI model trained on broad data, capable of being adapted for many downstream tasks. The model that produces a chatbot’s output is, in the cases FAIR cites, a foundation model.

Large language model (LLM)

A foundation model trained on natural-language text. The technical basis for most general-purpose AI chatbots in 2026.

Generative AI

AI services that produce new content (text, images, video, audio) on prompt, as distinct from AI services that classify or score existing data.

Automated decision-making (“ADM”)

A decision produced by automated processing without meaningful human involvement. Engaged by article 22 UK GDPR where the decision has legal or similarly significant effects.

AI-generated CSAM

Child sexual abuse material produced by AI models. The IWF’s 2026 report classified 65 per cent of identified AI-generated CSAM videos in 2025 as Category A, the most severe category.

Nudification / synthetic intimate imagery

The use of AI to generate sexually explicit imagery of a real person without consent. Criminalised in part by the Online Safety Act 2023 and subsequent legislation.

DMARC, SPF, DKIM

The three email-authentication standards. SPF lists the servers authorised to send mail for a domain; DKIM cryptographically signs outbound mail; DMARC tells receiving servers what to do with messages that fail SPF or DKIM, and provides reporting. FAIR’s domain runs all three.

DNSSEC

Domain Name System Security Extensions. Cryptographically signs DNS records so resolvers can detect tampering. Enabled on airights.org.uk.

Charity governance

CC9 (Charity Commission guidance)

The Charity Commission’s statutory guidance Speaking out: guidance on campaigning and political activity by charities. Permits campaigning and political activity in furtherance of a charity’s purposes, provided that political activity does not become the reason for the charity’s existence.

Charitable Incorporated Organisation (“CIO”)

A corporate form for charities under the Charities Act 2011. FAIR is constituted as a CIO (registration with the Charity Commission pending).

Charitable purposes

The purposes a charity is established to advance, drawn from the statutory list at section 3 of the Charities Act 2011. FAIR’s purposes are the advancement of education in respect of artificial intelligence and the advancement of human rights as engaged by artificial intelligence.

Public benefit

The statutory test that a charity must be established for the benefit of the public, not a closed class. Applied in the round, having regard to the Charity Commission’s guidance.

Trustees

The persons who have general control and management of the administration of a charity, with statutory and fiduciary duties. A CIO requires at least three trustees. Recruitment for FAIR’s founding trustees is open.

DPIA (Data Protection Impact Assessment)

The structured assessment required by article 35 UK GDPR before processing likely to result in a high risk to individuals’ rights and freedoms. Applies to certain AI deployments.

FAIR-specific

Speakers framework

FAIR’s proposed legal architecture, set out in Briefing No. 1, under which AI services that produce content are treated, for liability purposes, as legal speakers rather than as hosts of user-generated content.

Hosts framework

The dominant architecture of online-safety law as at 2026, including the Online Safety Act 2023, treating platforms as intermediaries for the speech of their users. FAIR argues this framework is inadequate to AI services that themselves speak.

Independence statement

FAIR’s standing commitment, repeated in each submission and briefing, that FAIR is independent of Government, of any political party, and of the AI industry, and does not accept funding from AI developers or deployers, or from any organisation whose commercial interests would be materially affected by FAIR’s public positions.

Glossary maintained by The Foundation for Artificial Intelligence Rights (FAIR), a UK organisation (registration pending). Suggestions and corrections welcome at info@airights.org.uk. See FAIR’s Charitable Purposes and Briefings.