The Digital Omnibus, in plain terms
The European Commission proposed the "Digital Omnibus on AI" on 19 November 2025, aiming to simplify parts of the AI Act (Regulation (EU) 2024/1689) after it became clear that the regulatory infrastructure — harmonised standards, national market-surveillance bodies — wasn't going to be ready in time for the original August 2026 deadline. After trilogue negotiations, the European Parliament adopted the agreed text on 16 June 2026 (423 votes in favor, 57 against, 174 abstentions), and the Council of the EU gave its final green light on 29 June 2026, closing the legislative file. The act was signed on 8 July 2026 and, as of this writing, is awaiting publication in the EU's Official Journal, where it enters into force on the third day after publication.
What it actually changes is narrower than the "AI Act delayed" headlines suggest:
- Standalone high-risk AI systems (Annex III — recruitment tools, credit-scoring, law-enforcement and border-control systems, education and exam-scoring tools, and similar) move from 2 August 2026 to 2 December 2027.
- High-risk AI embedded in regulated products (Annex I — medical devices, machinery, connected vehicles) move from 2 August 2027 to 2 August 2028.
- A new ban on AI systems that generate non-consensual intimate imagery ("nudifiers") and child sexual abuse material was added to Article 5, with a transitional period until 2 December 2026.
- The Commission's AI Office picks up broader supervisory powers, including over AI systems built on general-purpose AI models from the same provider.
None of that touches Article 50. It was never part of what the Omnibus delays.
Still on schedule for August 2, 2026 vs. actually delayed
| Still on schedule — 2 August 2026 | Delayed by the Digital Omnibus |
|---|---|
| Article 50(1) — disclosure that a person is interacting with an AI system (chatbots, virtual assistants) | Annex III high-risk AI (hiring, credit scoring, law enforcement, education) → now 2 December 2027 |
| Article 50(3) — emotion-recognition and biometric-categorisation systems must inform people exposed to them | Annex I high-risk AI embedded in regulated products (medical devices, machinery, vehicles) → now 2 August 2028 |
| Article 50(4) — labelling of AI-generated or manipulated content (deepfakes, synthetic media) | Article 50(2) machine-readable watermarking — but only for AI systems already on the market before 2 Aug 2026 (grace period to 2 Dec 2026) |
| The Commission's AI Office gains full investigative and fining powers over general-purpose AI (GPAI) providers | New Article 5 ban on non-consensual intimate imagery / CSAM — itself added by the Omnibus, transitional period to 2 Dec 2026 |
In other words: the only part of Article 50 that moves at all is the narrow watermarking rule in Article 50(2), and only for systems already deployed before the deadline. If your site has an AI chatbot or publishes AI-generated content, the parts of Article 50 that affect you directly are unaffected.
What Article 50 requires if your site has an AI chat widget
Most businesses reading this don't operate a high-risk AI system — they run a chat widget, a support bot, or an AI-powered search assistant. For that category of site, Article 50 boils down to two practical questions:
- Is it obvious to a visitor that they're talking to an AI, not a person? If it's clearly labelled ("AI Assistant," a bot avatar, an opening line that says so), you may already be covered — the obligation only kicks in where it wouldn't otherwise be obvious from context. If your widget could plausibly be mistaken for a human agent, Article 50(1) requires a disclosure.
- Do you publish AI-generated or AI-manipulated content — images, audio, video or synthetic media — that could be mistaken for authentic, human-made content? Article 50(4) requires that content to be disclosed as artificially generated or manipulated, particularly where it resembles real people, places or events.
Neither obligation requires you to be a "high-risk" AI deployer. Article 50 is a transparency rule that applies regardless of how risky the underlying AI use is — which is exactly why it wasn't part of the high-risk delay.
A quick, honest self-check
- Chat widget or AI assistant on your site? If it reaches EU visitors, Article 50(1) likely applies to it.
- Not 100% obvious it's AI? Add one visible line — "You're chatting with an AI assistant" — near the first interaction. It costs nothing and removes any ambiguity.
- Publishing AI-generated images, audio, video, or synthetic voice content? Label it as AI-generated or AI-assisted where it could be mistaken for authentic material.
- Using an AI system for emotion recognition or biometric categorisation (uncommon for most sites, but check any AI-powered analytics or camera-based tools)? Article 50(3) requires informing the people exposed to it.
- Keep a dated record of when you added disclosures and labelling — useful evidence of a good-faith effort regardless of how enforcement develops.
See what our free scanner finds on your site
Abledly's Article 50 checker looks for AI chat widgets and possible AI-content signals on your site and tells you what disclosure Article 50 requires — free, no signup to see your exposure summary.
Run the free Article 50 scanner →Frequently asked questions
Was the EU AI Act delayed?
Not entirely. The Digital Omnibus on AI — adopted by the European Parliament on 16 June 2026 and the Council of the EU on 29 June 2026 — postpones the compliance deadlines for high-risk AI systems to 2 December 2027 (standalone, Annex III) and 2 August 2028 (embedded in products, Annex I). It does not delay Article 50's transparency obligations, which still take effect 2 August 2026 as originally written into the AI Act in 2024.
What is the EU Digital Omnibus on AI?
A package of amendments to the AI Act (Regulation (EU) 2024/1689), proposed by the European Commission on 19 November 2025. Parliament approved the agreed text on 16 June 2026 (423 votes in favor), the Council gave final approval on 29 June 2026, and the act was signed 8 July 2026. As of this writing it's awaiting Official Journal publication, where it takes effect on the third day after publication.
What still applies on August 2, 2026?
Article 50: disclosure that a person is interacting with an AI system such as a chatbot (50(1)), obligations for emotion-recognition and biometric-categorisation systems (50(3)), and labelling of AI-generated or manipulated content like deepfakes (50(4)). The Commission's AI Office also gains full investigative and fining powers over general-purpose AI providers on this date. Only Article 50(2), machine-readable watermarking for systems already on the market before 2 August 2026, gets a grace period to 2 December 2026.
Do I need to disclose that my chatbot is AI?
If your website has a chatbot or AI assistant interacting with visitors in the EU, Article 50(1) generally requires making clear they're dealing with an AI system rather than a human, unless that's already obvious from context. A simple, visible disclosure near the first interaction is the low-risk way to satisfy this. General information, not legal advice for your specific case.
Does Article 50 apply to businesses outside the EU?
The AI Act's territorial scope is based on where an AI system's output is used, not just where the provider or deployer is based — similar in logic to GDPR and the European Accessibility Act. A US or Canadian business whose chatbot or AI content reaches EU users can be in scope without an EU office. Talk to a lawyer about your specific situation.
Is this page legal advice?
No. It's general information about a fast-moving regulatory change, sourced throughout, and not a substitute for advice from a qualified lawyer who knows your business and jurisdiction.
Sources
- Council of the EU — "Artificial intelligence: Council gives final green light to simplify and streamline rules" (29 June 2026 press release)
- Gibson Dunn — "EU AI Act Omnibus Agreement — Postponed High-Risk Deadlines and Other Key Changes" (client alert)
- Regulation (EU) 2024/1689 (the EU AI Act), Article 50 — original transparency obligations
Find out what's actually on your site before August 2
Free scan for AI chat widgets and possible AI-content signals, with the specific Article 50 disclosure each one needs.
Run your free Article 50 scan →