EU AI Act Timeline: Key Deadlines 2025–2027 for Website Owners
The EU AI Act (Regulation 2024/1689) entered into force on August 1, 2024, but its obligations take effect in stages over three years. Most website owners and small businesses are primarily affected by the August 2, 2026 transparency deadline — but the full timeline matters for planning. This article gives you the complete picture.
Bottom line: For the vast majority of websites — those using chatbots, AI-generated content, or AI images — the only date that matters right now is August 2, 2026. High-risk AI system obligations are separate and apply mainly to regulated sectors and enterprise software, not typical WordPress sites.
Full EU AI Act timeline
EU AI Act enters into force
Regulation 2024/1689 published in the Official Journal. 24-month application clock begins.
Prohibited AI practices apply (Chapter II)
The ban on prohibited AI systems took effect. Includes social scoring, real-time biometric surveillance in public, and subliminal manipulation. This does not affect ordinary websites.
General-purpose AI (GPAI) model obligations apply
Obligations for GPAI model providers (AI model developers like OpenAI, Google, Meta). Deployers — website owners using these models via APIs or plugins — are not directly subject to GPAI obligations.
Article 50 transparency obligations apply — the main deadline for websites
All deployers using AI chatbots or generating synthetic content must comply. Specifically: chatbot disclosure before/at start of conversation; labelling of AI-generated text in public interest contexts; disclosure of deepfake images/audio. This is the critical date for most website owners.
Machine-readable marking for AI-generated content (grandfathering)
Pre-existing generative AI systems (those already on the market before Aug 2026) have until December 2, 2026 to implement machine-readable watermarking/marking. This primarily affects AI tool providers, not deployers using those tools on their websites.
High-risk AI system obligations fully apply (Annex III systems)
The strictest obligations — conformity assessments, technical documentation, human oversight, registration in the EU database — apply to Annex III high-risk AI systems. These include AI used in critical infrastructure, employment decisions, education, law enforcement, and healthcare. Standard websites and chatbots are not high-risk AI systems.
Digital Omnibus Act amendments — high-risk obligations delayed for some
EU institutions reached a provisional agreement in May 2026 (the "Digital Omnibus") to delay certain high-risk AI obligations — Annex III high-risk applications are deferred to 2 December 2027. The August 2, 2026 Article 50 transparency deadline was explicitly not delayed and remains unchanged.
What this means for your website: a practical filter
- No AI features at all → No AI Act obligations. Update your privacy policy if you use third-party AI tools in the background.
- AI chatbot → August 2, 2026: add the chatbot disclosure. Done.
- AI-generated product descriptions or blog content → August 2, 2026: the labelling obligation only applies if your content addresses matters of general public interest (e.g., public health, political opinion). Commercial product descriptions are not in scope.
- AI-generated images or video → August 2, 2026: label deepfakes (realistic synthetic representations of real people). General AI imagery without a real person depicted is not a deepfake under the Act.
- High-risk AI use (e.g., AI-assisted hiring, creditworthiness assessment, student grading) → Consult a specialist. The 2027 deadline and conformity assessment obligations apply.
What the Digital Omnibus changed (May 2026)
In May 2026, EU institutions reached a provisional agreement (the Digital Omnibus) to amend the AI Act, deferring certain high-risk (Annex III) obligations to 2 December 2027. This remains provisional until formally adopted and published in the Official Journal. The key point for most website owners: Article 50 transparency obligations were not delayed. The August 2, 2026 deadline for chatbot disclosure and AI content labelling stands unchanged.
Your action plan before August 2, 2026
- Check whether your site uses an AI chatbot → if yes, add the disclosure to the welcome message
- Check whether your AI-generated text appears in public-interest contexts → if yes, add a label
- Check whether you publish deepfake imagery → if yes, add visible disclosure
- Add or update your AI transparency statement page
- Run the Legibright checker on your URL to get your compliance score
This article is intended for general information purposes only and does not constitute legal advice. For advice specific to your situation, consult a qualified legal professional. Dates reflect the regulation as published and amended as of mid-2026; always verify against the current text of the regulation.