Part of the AI Regulation News hub.
EU AI Act Enforcement and Article 50 Start
From 2 August 2026, the Commission and national authorities begin enforcing applicable AI Act rules, and new Article 50 transparency requirements start applying to certain interactive systems and AI-generated or altered content.
Bottom line: The August 2 milestone moves specified AI Act requirements from guidance and preparation into enforcement and application.
Who this affects: Operators of covered interactive AI systems and organisations producing AI-generated or altered content.
What changed: Enforcement of applicable provisions begins, and the Commission says new transparency rules under Article 50 start applying.
What did not change: The Commission describes a defined set of transparency requirements, not a rule for every AI system.
- Instrument
- Regulation (EU) 2024/1689, the Artificial Intelligence Act.
- Authority
- European Commission AI Office and national authorities.
- Jurisdiction
- European Union.
- Status
- Enforcement and specified transparency requirements begin on 2 August 2026.
- Primary source
- European Commission notice, 31 July 2026
- Last verified
- 4 August 2026.
What starts on 2 August?
The Commission says its AI Office and national authorities will begin enforcing applicable provisions of the AI Act from 2 August 2026. On that same date, new transparency rules under Article 50 start applying.
This is a status update about the operative date. It is distinct from the Commission's May guidance explaining Article 50 transparency obligations.
Which transparency measures does the Commission identify?
The Commission says chatbots and other interactive AI systems covered by the rules must tell users they are interacting with AI rather than a human. It says deepfakes must be labelled, and AI-generated or altered content must carry machine-readable marks so it can be detected more easily.
The Commission frames these measures as tools to reduce deception and manipulation and help people make informed choices.
What should an EU-facing AI team do now?
Map every interactive AI touchpoint and the content it produces or alters. For each covered system, identify the user notice, labelling process and machine-readable marking method that the team relies on.
Keep the evidence supporting those controls with the owner responsible for the relevant system. This page does not decide whether a particular system is covered. That assessment depends on the AI Act and the facts of its use.
| Question | Current answer |
|---|---|
| When does enforcement begin? | The Commission says its AI Office and national authorities begin enforcing the AI Act on 2 August 2026. |
| What must covered chatbots tell users? | They must tell users they are interacting with AI rather than a human. |
| What must happen to deepfakes? | The Commission says they must be labelled. |
Do not treat the August 2 milestone as just another guidance release. Confirm whether each system falls within the transparency rules, then document the relevant user notice, label and machine-readable mark.
Primary source: https://digital-strategy.ec.europa.eu/en/news/commission-starts-enforcing-ai-act-rules-and-new-transparency-requirements-2-august
FAQ
When did EU AI Act enforcement begin?
The Commission says its AI Office and national authorities begin enforcing the AI Act from 2 August 2026.
Do transparency rules apply to every AI system?
No. The Commission identifies certain interactive AI systems and AI-generated or altered content. Coverage depends on the AI Act and the facts of use.
What is the Article 50 practical change?
Covered interactive systems must tell users they are dealing with AI. The Commission also says deepfakes must be labelled and AI-generated or altered content must carry machine-readable marks.