AI Policy Wiki
Dashboard
⚲
AI Policy Wiki › People & Organizations

European Parliament

medium confidence · updated 2026-09-24

The directly elected legislative body of the European Union; co-legislator of the EU AI Act, the Digital Markets Act and the GDPR. Negotiated the 2026 Digital Omnibus on AI with the Council, held a May 2026 committee hearing on Anthropic's Mythos, and in July 2026 was reported to be rolling out an internal 'EPGenAI Hub' giving MEPs and staff access to Meta, OpenAI, Anthropic, and Mistral models.

Contents

The European Parliament is the directly elected legislative body of the European Union. Together with the Council of the European Union it adopts EU legislation, including the EU AI Act, and its committees conduct oversight of the European Commission's implementation work, including the EU AI Office.

Role in AI legislation

The Parliament served as co-legislator on the EU AI Act, the Digital Markets Act, and the Digital Services Act, and continues to shape amendment activity such as the Digital Omnibus on AI (see EU AI Act (Regulation 2024/1689) for the Act's reform timeline). The AI Act was adopted as Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 (EU AI Act (Regulation 2024/1689)); the Digital Markets Act was adopted by the Parliament and the Council on September 14, 2022 as Regulation (EU) 2022/1925 (Regulation (EU) 2022/1925 — Digital Markets Act (primary text)); and the General Data Protection Regulation was adopted as Regulation (EU) 2016/679 on 27 April 2016 (General Data Protection Regulation (GDPR)).

Digital Omnibus on AI

The Parliament negotiated the 2026 Digital Omnibus on AI, the package that postponed the AI Act's high-risk obligations, with the Council of the EU. After 12 hours of overnight talks beginning 28 April 2026, the two institutions failed to reach a common position on the reform package (Source: reuters.com), and a further trilogue on delaying the high-risk provisions broke down on 30 April / 1 May over whether machinery and medical-device AI should fall under sectoral law or the AI Act's high-risk regime (Source: artificialintelligenceact.substack.com). A political agreement between the Parliament and the Council became public on 20 May 2026. It moved the high-risk rules to 2 December 2027 and product-embedded systems to 2 August 2028, and added a prohibition on AI systems that generate non-consensual sexually explicit content, including "nudification" apps (Source: artificialintelligenceact.substack.com). The Council gave the package final approval on 29 June 2026 (Source: consilium.europa.eu).

Digital Markets Act enforcement

On 30 April 2026 the Parliament adopted a resolution on Digital Markets Act enforcement (2026/2596(RSP)) calling for closer scrutiny of AI-driven search tools, naming Google AI Overviews, Gemini, Apple Siri, Meta's WhatsApp AI, Amazon Rufus and Microsoft Copilot, and for complementarity between the DMA, the Data Act and the AI Act (Source: techieray.substack.com).

Protection of minors online

In a report of November 2025 on the protection of minors online, the Parliament called for a harmonised European digital age limit of 16 for access to social media, video-sharing platforms and AI companions unless parents or guardians authorise otherwise, and a harmonised age of 13 below which no minor could access social media. It also called for further legislation on age-appropriate design and safety by design and for a European approach to age assurance. The Commission cites the report in the explanatory memorandum of its EU KIDS Act proposal of 17 September 2026, which sets the autonomous-account age at 15 with guardian-set accounts from 13, and which requires the Parliament's agreement with the Council to become law (Proposal for an EU KIDS Act, COM(2026) 681 — Source Summary).

Oversight of frontier AI

The Parliament's internal market committee invited Anthropic to a hearing on 6 May 2026 on its Mythos model, with Sarah Heck of Anthropic, Commission Executive Vice-President Henna Virkkunen, AI Office head Lucilla Sioli and representatives of ENISA expected to attend (Source: artificialintelligenceact.substack.com). In a letter disclosed the week of 18 May 2026, 30 Members of the European Parliament from six political groups warned Virkkunen that EU cybersecurity law is "ill-equipped" for tools such as Mythos and urged that ENISA be given model access (Source: artificialintelligenceact.substack.com).

The Parliament is also the forum for the Commission President's annual State of the Union address. In the address delivered in Strasbourg on September 16, 2026, Ursula von der Leyen endorsed calls from frontier-laboratory leaders to "pace the frontier" and said she would invite the main frontier labs to discuss how to support those efforts (Source: reuters.com). See Coordinated slowdown proposals compared.

Internal AI adoption

Politico reported on July 20, 2026 that the Parliament plans to roll out an "EPGenAI Hub" giving MEPs and staff access to models from Meta, OpenAI, Anthropic, and Mistral as early as September 2026 (Source: politico.eu). The deployment places the institution among government bodies adopting the same class of systems its legislation regulates.

Civil liability for AI

On 20 October 2020 the Parliament adopted a legislative own-initiative resolution under Article 225 TFEU (2020/2014(INL)) asking the Commission to propose a civil liability regime for AI based on Article 114 TFEU. The resolution suggested a limited strict-liability regime for certain AI-enabled technologies together with an eased burden of proof under fault-based rules. The Commission's September 2022 AI Liability Directive proposal took up the burden-of-proof measures and deferred strict liability and mandatory insurance to a later review; it cited the Parliament's resolution of 3 May 2022 on artificial intelligence in a digital age as sharing its concerns (Proposal for an AI Liability Directive, COM(2022) 496 — Source Summary). After the Commission withdrew the proposal in 2025, four MEPs, Kim van Sparrentak, Axel Voss, Brando Benifei and Michael McNamara, asked the Commission in a note dated 16 September 2026 to introduce an "AI Liability Act" (Source: politico.eu).

Relationships

Categories: People & Organizations | eu | government | legislation | deployer
What links here (10)