cd /news/ai-policy/labelling-ai-written-articles-and-sy… · home › topics › ai-policy › article
[ARTICLE · art-151693] src=dev.to ↗ pub= topic=ai-policy verified=true sentiment=· neutral

Labelling AI-written articles and synthetic media under Article 50

A new guide explains how deployers of generative AI must comply with Article 50(4) of the EU AI Act, which requires disclosure for AI-written text on matters of public interest unless a human has reviewed the content and a named person or organisation holds editorial responsibility. It also notes that realistic synthetic images, audio and video require a visible label regardless of human review, and that the separate marking duty under Article 50(2) falls on the provider of the generation tool. The guide recommends a plain disclosure line such as "This article was generated by AI" placed at or before first publication.

by read2 min views3 publishedOct 17, 2026

This guide was drafted by an AI agent from the primary sources linked at the bottom, and has not been reviewed by a lawyer. It is general information, not legal advice. The original is at https://mdodeja.github.io/eu-ai-act-article-50-guide/guides/labeling-ai-written-articles.html.

Short answer. AI-written text on matters of public interest needs a disclosure unless a human reviewed it and a named person or organisation holds editorial responsibility. Realistic synthetic images, audio and video need a visible label either way.

Article 50(4) puts two duties on deployers of generative AI, and agencies often mix them up.

The marking duty in Article 50(2) is separate again, and it belongs to the provider of the generation tool. The marking guide covers it.

For the text duty, Article 50(4) has an exception. You do not have to disclose AI generation if the content has been through human review or editorial control, and a natural or legal person holds editorial responsibility for its publication.

That gives an agency a practical route for AI-drafted articles:

If no human is in the loop, label the piece. A plain line works: "This article was generated by AI." It should appear at or before first publication, where readers can see it.

The Act does not spell out how deep a review has to be. The Commission's guidelines on Article 50 address the scope of these duties, so read them for your use case, and treat a skim as too little.

Ordinary marketing copy, such as product pages and category descriptions, is generally not text published to inform the public on matters of public interest. Content that reads like news, commentary, or civic information is. When a piece could be taken for either, handle it as public-interest text.

For realistic synthetic media, a metadata mark alone is not enough. The disclosure has to be visible to the person looking at it.

The Commission's Code of Practice includes a set of icons that deployers may use for labelling.

Pipeline:        ____________   Client: ____________   Date: ________
Classification:  fully AI-generated / AI-assisted / assistive only
50(2) marking:   tool-native metadata kept / C2PA added / not applicable (why)
50(4) media:     visible label applied / not applicable
50(4) text:      disclosed / editorial responsibility held by NAME / not applicable
Evidence:        links or screenshots
── more in #ai-policy 4 stories · sorted by recency
── more on @european union 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/labelling-ai-written…] indexed:0 read:2min 2026-10-17 · —