Article 50 duty · Art. 50(4)

Art. 50(4) — Disclose that the content has been artificially generated or manipulated

Who this falls on: The deployer.
Applies from: 2 August 2026.

What triggers it

Deepfakes (image, audio or video content that is artificially generated or manipulated); AI-generated or manipulated text published with the purpose of informing the public on matters of public interest

What you must do

Disclose that the content has been artificially generated or manipulated

The exception

Law-enforcement carve-out

Does not apply where the use is authorised by law to detect, prevent, investigate or prosecute criminal offences.

The dates that apply

Art. 50 applies from2 August 2026
Art. 50(2) marking transition (systems already on the market)2 December 2026
Annex III high-risk2 December 2027
High-risk embedded in regulated products2 August 2028

Regulation (EU) 2026/1744 (Digital Omnibus on AI) of 8 July 2026, OJ 24 July 2026, in force 27 July 2026 (ELI: http://data.europa.eu/eli/reg/2026/1744/oj), moved the high-risk dates to 2027-12-02 (Annex III) and 2028-08-02 (Annex I); it did NOT delay Article 50's application date, which remains 2026-08-02.

How this shows up in a questionnaire

Buyers rarely quote the paragraph number. They ask it as a control: whether users are told they are dealing with an AI system, whether generated output is marked, whether your disclosure fires before the first interaction. An answer that names the paragraph, the role and the date reads as prepared. One that says “we comply with the EU AI Act” reads as boilerplate and invites a follow-up round.

All five Article 50 duties

This duty is a question on a live deal’s questionnaire?

Send the questionnaire. First 3 answers free within 24h, full delivery in 48h for $490 flat up to 60 questions, paid after delivery, late means free.

Send your questionnaire →