EU AI Act transparency rules: what now applies
The EU AI Act became broadly applicable on August 2, 2026. See which transparency duties are in force, what the AI Omnibus delayed, and the next deadlines.
Short answer
Since August 2, 2026, the EU AI Act applies in general, including the Article 50 transparency duties for chatbots, AI-generated content and deepfakes. The AI Omnibus pushed most high-risk obligations to December 2, 2027 and August 2, 2028, but left the transparency rules in place.
Key takeaways
- Chatbot providers must tell people they are talking to an AI system unless that is obvious from context.
- Generative AI systems already on the market before August 2, 2026 have until December 2, 2026 to add machine-readable marking of their output.
- High-risk rules now start on December 2, 2027 (Annex III uses) and August 2, 2028 (AI in regulated products).
On August 2, 2026, the EU AI Act became applicable in general, two years after it entered into force on August 1, 2024. That date switched on the Act's transparency rules for chatbots and AI-generated content, and handed enforcement to the EU AI Office and national authorities. A separate law, the AI Omnibus, moved several later deadlines just days earlier. Here is where things stand as of November 2026.
What changed on August 2, 2026?
The AI Act became broadly applicable, and its Article 50 transparency obligations became enforceable across the EU.
The European Commission's AI Act page lists August 2, 2026 as the date the regulation applies, with some exceptions, and the point where the AI Office and Member State authorities took on supervision and enforcement. Earlier phases had already started: banned practices and AI literacy duties on February 2, 2025, and rules for general-purpose AI models on August 2, 2025.
The Article 50 duties, as summarized by law firm Goodwin, cover four situations:
- AI systems that interact with people must make clear that the user is dealing with an AI, unless it is obvious.
- Systems that generate synthetic audio, images, video or text must mark their output in a machine-readable, detectable format.
- Deployers must disclose deepfakes that depict real people, places or events.
- Deployers of emotion recognition or biometric categorization systems must inform the people exposed to them.
What did the AI Omnibus delay?
The Omnibus moved high-risk obligations to late 2027 and 2028 but did not change the substance of the transparency rules.
The Omnibus was proposed on November 19, 2025, reached political agreement on May 7, 2026 and entered into force on July 27, 2026, according to the Commission. Goodwin identifies it as Regulation (EU) 2026/1744.
| Milestone | Date |
|---|---|
| AI Act enters into force | August 1, 2024 |
| Prohibited practices and AI literacy | February 2, 2025 |
| General-purpose AI model rules | August 2, 2025 |
| AI Omnibus enters into force | July 27, 2026 |
| General application, transparency rules | August 2, 2026 |
| Marking deadline for systems already on the market | December 2, 2026 |
| New ban on AI-generated non-consensual sexual content and CSAM | December 2026 |
| High-risk rules, Annex III uses | December 2, 2027 |
| High-risk rules, Annex I products | August 2, 2028 |
Annex III covers uses such as biometrics, education, employment and migration. Annex I covers AI built into products that already fall under EU product safety laws.
What changes for users?
People in the EU should see clearer labels when they talk to a chatbot or view AI-generated media, with machine-readable marks becoming mandatory for older systems by December 2, 2026.
In practice that means disclosure notices in chat apps and customer-service bots, and labels or metadata on generated images and video. The marking rule targets detectability by software, so a watermark or embedded tag may not be visible to the eye. Labels do not make an answer correct; a chatbot can still invent facts, which our explainer on AI hallucinations walks through.
What happens to companies that break the rules?
Fines for Article 50 breaches can reach 15 million euros or 3% of global annual turnover, whichever is higher.
The new prohibitions arriving in December 2026 carry the Act's top tier of up to 35 million euros or 7%. The Commission also published guidelines on transparency obligations on July 20, 2026, and the first harmonized standard supporting the Act appeared in the Official Journal on October 19, 2026.
What is still unclear?
How strictly national authorities will police the new labels, and how marking will work across platforms, is still taking shape.
Enforcement sits mostly with national market surveillance authorities, so early practice may differ between countries. The Commission's own AI Act page notes that Executive Vice-President Virkkunen updated the Council and Parliament on simplification and enforcement on November 4, 2026. If you use AI assistants daily, the disclosure labels are a good prompt to check what data you share; our guide to AI chatbot privacy settings shows where those controls live, and how large language models work explains what sits behind the label.
Frequently asked questions
Did the AI Omnibus delay the transparency rules?
No. The Omnibus, Regulation (EU) 2026/1744, entered into force on July 27, 2026 and moved high-risk deadlines, but the Article 50 transparency obligations still apply from August 2, 2026. The only grace period is for machine-readable marking by generative systems already on the market, which runs to December 2, 2026.
Who enforces the AI Act transparency rules?
The European Commission says the AI Office and Member State authorities took on implementation, supervision and enforcement from August 2, 2026. National market surveillance authorities handle most Article 50 cases.
What are the fines for breaking Article 50?
Under Article 99(4), fines can reach 15 million euros or 3% of worldwide annual turnover, whichever is higher. The new prohibitions that apply from December 2, 2026 carry fines of up to 35 million euros or 7%.
Does this affect people using chatbots outside the EU?
The obligations apply to AI systems offered or used in the EU. Users elsewhere may still see the same disclosure labels if a provider applies one design worldwide, but the law itself covers the EU market.
Sources
- European Commission: AI Act, regulatory framework for AI
- Goodwin: EU AI Act transparency obligations are now in force
Research-based: Technorat does not claim hands-on testing. Facts as stated by the sources above, as of November 2026. See how we research. Spotted an error? contact@technorat.com