Civil society calls on the EU to put fundamental rights first in the AI Act

Civil society calls on the EU to put fundamental rights first in the AI Act — published under the document-side title An EU Artificial Intelligence Act for Fundamental Rights — is the 30 November 2021 joint civil-society statement coordinated by European Digital Rights (EDRi), signed by 120 organisations at publication (EDRi plus 119 civil-society co-signatories per the convener's own tally; AlgorithmWatch's parallel co-signatory announcement records the count as 115 and the subsequent reference cited in AlgorithmWatch's corpus entry as 123 reflecting post-launch additions), and co-drafted by a stable nucleus of European digital-rights organisations: EDRi, Access Now, the Panoptykon Foundation, epicenter.works, AlgorithmWatch, the European Disability Forum, Bits of Freedom, Fair Trials, PICUM, and ANEC. Published seven months into the AI Act legislative cycle — between the European Commission's 21 April 2021 proposal and the European Parliament's opening of co-rapporteur drafting — the statement is the foundational coalition manifesto under which the in-corpus EDRi-coordinated EU AI Act civil-society coalition for fundamental rights (2021–2024) operated through the trilogue endgame and into the implementation phase.

What the statement asks for

The statement organised the coalition's response to the Commission's proposal around two registers: a seven-item prohibition list and a nine-item set of architectural reforms. The seven prohibitions were social scoring; remote biometric identification in publicly accessible spaces (by all actors, not only law enforcement); emotion recognition; discriminatory biometric categorisation; AI physiognomy; systems predicting future criminal activity; and profiling and risk-assessment systems in migration contexts. The architectural reforms were a cohesive, flexible, and future-proof risk-categorisation methodology with robust update mechanisms; a comprehensive prohibition list on AI systems posing unacceptable rights risks; a mandatory obligation on users (not only providers) of high-risk AI systems to conduct and publish a fundamental-rights impact assessment; consistent public transparency through registration of high-risk system deployments in publicly accessible databases; meaningful rights and redress for affected individuals, including a right to an explanation for decisions taken with the assistance of AI systems; accessibility obligations throughout the AI lifecycle; sustainability and environmental protections with public disclosure of resource consumption and emissions; improved future-proof standards with limited harmonisation and civil-society participation in the standards-setting process; and comprehensive enforcement adequately resourced and centred on fundamental-rights protection. The statement's underlying critique of the Commission's 21 April 2021 draft was that its obligations sat predominantly on providers rather than users, that its list-based high-risk architecture was biased toward provider self-classification, and that its biometric-identification restrictions were narrowly bounded to law-enforcement use rather than extending across all actors in publicly accessible spaces.

The coalition behind the statement

The named lead-drafting circle — EDRi, Access Now, the Panoptykon Foundation, epicenter.works, AlgorithmWatch, the European Disability Forum, Bits of Freedom, Fair Trials, PICUM, and ANEC — is the stable nucleus of European digital-rights, disability-rights, fair-trials, migration-rights, and consumer-standardisation organisations whose national bases together cover Belgium (EDRi secretariat, EDF), Poland (Panoptykon), Austria (epicenter.works), Germany and Switzerland (AlgorithmWatch), the Netherlands (Bits of Freedom), and a Brussels presence on the migration-rights (PICUM) and consumer-standardisation (ANEC) sides. The broader signatory roster of 120 organisations at publication extended the coalition into national digital-rights organisations across the European Union, international rights organisations including Amnesty International and Privacy International, race-equality and migrants-rights organisations including Statewatch and the European Network Against Racism, and a substantial disability-rights tail anchored on the European Disability Forum's network. The Brussels secretariat for the drafting and co-ordination work was supplied by EDRi, whose role as the European digital-rights field's convening table is treated elsewhere in the corpus as the principal civil-society co-ordination point on the AI Act file. EDRi's Ella Jakubowska, who became the coalition's most consistent public spokesperson on the biometric-identification track through the subsequent legislative cycle, anchors the convener side of the statement on EDRi's own Brussels staff.

The statement's downstream organising trajectory

The 30 November 2021 statement is the document the EU AI Act civil-society coalition carried as its foundational position through the rest of the AI Act legislative cycle. The European Parliament's two co-rapporteurs subsequently drew on the statement's seven-prohibition list and on its user-side fundamental-rights impact-assessment demand as they assembled the Parliament's negotiating mandate, which was adopted on 14 June 2023. As the Act entered final trilogue negotiations in summer 2023, the same EDRi-coordinated coalition published a follow-on 12 July 2023 trilogue statement titled Civil society calls on EU to protect people's rights in the AI Act 'trilogue' negotiations, signed by 150 organisations and co-drafted with a partly overlapping nucleus that added Amnesty International, the European Center for Not-for-Profit Law, the Irish Council for Civil Liberties, Homo Digitalis, and the European Federation of National Trade Unions — a follow-on artefact that operated explicitly on the foundation the November 2021 manifesto had set. As the trilogue entered its closing stage, the coalition's 8 December 2023 "Don't Trade Away Our Rights!" open letter, signed by 70 organisations and 34 expert individuals — including EDRi, Amnesty International, Privacy International, Human Rights Watch, and Access Now — pressed the same seven-prohibition position list against the law-enforcement and migration-context carve-outs and the national-security exemption that emerging Council amendments threatened to introduce. Political agreement was reached on 9 December 2023, the European Parliament adopted the final text on 13 March 2024 by 523 votes in favour, the Council approved it on 21 May 2024, and the AI Act entered into force on 1 August 2024 — with full application set for 2 August 2026. The substantive trace of the November 2021 manifesto in the adopted text is identifiable: mandatory user-side fundamental-rights impact assessments, the partial bans on emotion recognition (in workplaces and educational settings) and on AI physiognomy and discriminatory biometric categorisation, social-scoring prohibition for public authorities, and the public-register obligations all survived the trilogue endgame, while the in-principle prohibition on remote biometric identification arrived with the law-enforcement and migration-context carve-outs the coalition had warned against.

Position in the corpus

Within the corpus, this is the foundational coalition-manifesto publication-side artefact on the EU AI Act file — the document under which the EDRi-coordinated coalition operated from 30 November 2021 through to the Act's August 2024 entry into force. It pairs with Automating Society Report 2020 on the Continental European side of the publications slate: the AlgorithmWatch / Bertelsmann Stiftung report supplied the recurring empirical evidence base on which European civil-society AI Act advocacy was built, and this statement supplied the coalition political-policy manifesto on which the EU institutional response was pressed — together they form the corpus's two anchoring Continental European AI-policy publication-side artefacts. As a publication type it fills the manifesto slot, structurally distinct from the corpus's only other coalition-letter publication (Artists 4 Safe AI, a ~125–300 individual-creative-signatory open letter to Governor Newsom against California SB 1047's veto, a structurally different category: ~300 individuals on a California state-bill veto versus 120+ civil-society organisations on the EU's horizontal AI law). It is the corpus's clearest single artefact on the surveillance / biometric-rights movement-area on the publication side — the area in which the EU AI Act coalition's named convener EDRi carries its public organising voice — and the publication-side anchor that the in-corpus EU AI Act civil-society coalition campaign body treats as its founding political document.