sebbi.pro
Legislation / European Union
EUROPEAN UNION

AI and the law in the European Union

In the EU, the Commission proposes laws and the Parliament and Council adopt them. Statuses are as of October 2026; the EUR-Lex link on each entry is the authoritative record.

Applying now

AI Act, Regulation (EU) 2024/1689applying

In force; applying in stages to 2028

The comprehensive AI law: prohibitions, high-risk requirements, transparency and general-purpose model duties.

Digital Omnibus on AI, Regulation (EU) 2026/1744in force

In force 27 July 2026

Moves high-risk deadlines to 2 December 2027 (Annex III) and 2 August 2028 (Annex I), adds a ban on nudification and child abuse material generation, eases rules for smaller firms.

Evidence: ai act/timeline

General Data Protection Regulation (EU) 2016/679applying

Applying since 25 May 2018

Lawful processing, accountability, data subject rights and safeguards for automated decisions (Article 22).

Evidence: checks/gdpr · schufa

Digital Services Act, Regulation (EU) 2022/2065applying

Applying since 17 February 2024

Statements of reasons for moderation (including automated), recommender transparency, systemic risk duties for very large platforms.

Digital Markets Act, Regulation (EU) 2022/1925applying

Applying since 2 May 2023

Obligations on designated gatekeepers, including limits on combining personal data and self-preferencing in ranking.

Data Act, Regulation (EU) 2023/2854applying

Applying since 12 September 2025

Access to connected-product data, switching between cloud providers, and smart contract safeguards (Article 36).

Evidence: data act 36

Cyber Resilience Act, Regulation (EU) 2024/2847applying

Reporting from 11 September 2026, main duties from 11 December 2027

Security by design for products with digital elements, and 24-hour reporting of actively exploited vulnerabilities.

NIS2 Directive (EU) 2022/2555applying

Transposition due 17 October 2024

Cybersecurity risk management and incident reporting for essential and important entities.

Evidence: checks/nis2

General Product Safety Regulation (EU) 2023/988applying

Applying since 13 December 2024

Safety duties for consumer products including connected ones, with accident reporting.

Evidence: gpsr

eIDAS 2, Regulation (EU) 2024/1183applying

In force 20 May 2024

European Digital Identity Wallets and trust services; electronic timestamps can't be refused as evidence for being electronic.

Evidence: eidas 41 · forever

European Accessibility Act, Directive (EU) 2019/882applying

Applying since 28 June 2025

Accessibility for e-commerce, banking, transport, e-books and communications.

Evidence: accessibility

Empowering Consumers for the Green Transition, Directive (EU) 2024/825applying

Applies from 27 September 2026

Bans generic and offset-based environmental claims; every claim needs evidence.

Evidence: green claims

Adopted, applying soon

Product Liability Directive (EU) 2024/2853adopted

Transposition due 9 December 2026

Software and AI are products; courts can presume a defect when the maker can't disclose evidence.

Machinery Regulation (EU) 2023/1230adopted

Applies from 20 January 2027

Safety requirements for machinery with self-evolving behaviour, including recording safety-related decisions.

Evidence: machinery

Pay Transparency Directive (EU) 2023/970adopted

Transposition due 7 June 2026

Objective, gender-neutral pay criteria and the employer's burden to prove no discrimination.

Platform Work Directive (EU) 2024/2831adopted

Transposition due 2 December 2026

Human oversight and review of algorithmic management of platform workers.

Evidence: platform work

Consumer Credit Directive (EU) 2023/2225adopted

Applies from 20 November 2026

Human intervention and explanation for automated creditworthiness assessments.

Evidence: credit ai

European Health Data Space, Regulation (EU) 2025/327adopted

In force March 2025, applying in stages

Patients' access to electronic health data and rules for secondary use of health data, including for training AI.

Withdrawn

AI Liability Directive (proposal)withdrawn

Withdrawn by the Commission in 2025

Would have eased proof of fault for AI harm. Withdrawn; AI harm is handled by national law and the new Product Liability Directive.

Evidence: liability

ePrivacy Regulation (proposal)withdrawn

Withdrawn by the Commission in 2025

Would have replaced the ePrivacy Directive on cookies and electronic communications. The Directive remains in force.

AI news
The VergeOpenAI announces three big science discoveries and a new way to work with academics33mThe VergeAnthropic is burning gas to feed the AI boom2hWiredWhy Anthropic’s IPO Prospectus Is Missing So Many Key Figures2hArs TechnicaApple releases iOS and macOS 27.2 with new Siri AI features, new Music logo4hBBC NewsAnthropic's blockbuster share listing - 5 surprising things we now know4hMIT Technology ReviewWill the real Anthropic please stand up6hThe VergeThe RealReal won’t let AI agents hog designer clothing6hThe VergeFive weird details from Anthropic’s IPO filing6hTechCrunchAnthropic’s IPO filing reveals its strategy for keeping enterprise customers hooked7hBBC NewsAnthropic warns of 'catastrophic risk' to humankind ahead of blockbuster share listing7hWiredThe Anthropic S-1 Filing Is Also an AI Doomsday Primer8hWiredAnthropic’s IPO Prospectus Reveals What’s Under the Hood9hTechCrunchAnthropic officially files for its public debut9hThe VergeKevin Roose is AGI-pilled. Why aren’t you?10hWiredMeet the AI Agents Trying to Be Real-Life Journalists14hMIT Technology ReviewThe focus on AI safety disasters ignores everyday harms15hWiredThis AI Radio Host Speaks Spanish and Backs Immigrants. It’s Winning Over Listeners15hTechCrunchGrok releases the first seven minutes of its AI-generated ‘Odyssey’1dTechCrunchAI delivers ‘zero financial payoff’ to most companies, Apollo economist says1dTechCrunchThe hottest AI wearables and gadgets you should know about1dBBC NewsUK to invest £100m in 'world first' AI Biosecurity Lab3dThe VergeOpenAI announces three big science discoveries and a new way to work with academics33mThe VergeAnthropic is burning gas to feed the AI boom2hWiredWhy Anthropic’s IPO Prospectus Is Missing So Many Key Figures2hArs TechnicaApple releases iOS and macOS 27.2 with new Siri AI features, new Music logo4hBBC NewsAnthropic's blockbuster share listing - 5 surprising things we now know4hMIT Technology ReviewWill the real Anthropic please stand up6hThe VergeThe RealReal won’t let AI agents hog designer clothing6hThe VergeFive weird details from Anthropic’s IPO filing6hTechCrunchAnthropic’s IPO filing reveals its strategy for keeping enterprise customers hooked7hBBC NewsAnthropic warns of 'catastrophic risk' to humankind ahead of blockbuster share listing7hWiredThe Anthropic S-1 Filing Is Also an AI Doomsday Primer8hWiredAnthropic’s IPO Prospectus Reveals What’s Under the Hood9hTechCrunchAnthropic officially files for its public debut9hThe VergeKevin Roose is AGI-pilled. Why aren’t you?10hWiredMeet the AI Agents Trying to Be Real-Life Journalists14hMIT Technology ReviewThe focus on AI safety disasters ignores everyday harms15hWiredThis AI Radio Host Speaks Spanish and Backs Immigrants. It’s Winning Over Listeners15hTechCrunchGrok releases the first seven minutes of its AI-generated ‘Odyssey’1dTechCrunchAI delivers ‘zero financial payoff’ to most companies, Apollo economist says1dTechCrunchThe hottest AI wearables and gadgets you should know about1dBBC NewsUK to invest £100m in 'world first' AI Biosecurity Lab3d