EU AI Act Article 50 Deadline: What Changes on 2 August 2026

Two days from now, on 2 August 2026, Article 50 of the EU AI Act stops being “the thing everyone’s been meaning to prepare for” and becomes an enforceable transparency obligation. If you publish AI-generated or AI-manipulated content professionally — text, video, audio, social posts — this is the date the disclosure duty actually bites. Here’s what that means in practice, and what it doesn’t.

What actually changes on 2 August

Nothing about the rule changes — Article 50(4) and (5) have read the same way since the AI Act was published in 2024. What changes on 2 August is that the obligation becomes applicable law, not a future date on a compliance roadmap. Deployers who publish AI-generated text on matters of public interest, or AI-generated/manipulated deepfake-style image, audio, or video, must disclose that clearly and distinguishably, at the point a reader or viewer first encounters the content — not in a footer, not in end credits, not behind a linked policy page.

If you’ve already read the full disclosure guide and adjusted your publishing workflow, this date changes nothing for you operationally. If you haven’t, this is the actual line: content published on or after 2 August without a compliant disclosure is now a live compliance gap, not a future one.

The one grace period that matters

There’s a genuine carve-out, and it’s easy to miss. The July 2026 “AI Omnibus” amendment (Regulation (EU) 2026/1744) gives providers of generative AI systems that were already on the market before 2 August 2026 until 2 December 2026 to bring the underlying content-marking mechanisms into compliance with Article 50(2) — the machine-readable marking obligation that sits on providers, not deployers.

Don’t read that as a personal extension. It’s a technical transition window for AI system providers to update their tooling, not a delay to your disclosure obligation as a deployer publishing content. If you’re publishing AI-assisted content on 2 August, your disclosure duty starts that day regardless of whether the tool you used has finished rolling out its own machine-readable markers.

What isn’t happening on 2 August

The other headline EU AI Act date most people have heard — high-risk AI system obligations under Annex III (recruitment, credit scoring, education, and similar) — is not what’s arriving this week. That deadline was pushed to 2 December 2027 by the same Omnibus amendment. If you’ve been putting off high-risk classification work because “the August deadline is coming,” you have more runway than the old date suggested — see the full deadline breakdown if that’s what you’re tracking.

What to check before Sunday

1. Audit anything you’re about to publish or have scheduled for 2 August onward — blog posts, social captions, video, podcast notes — and confirm the disclosure sits before the first line of substantive content, not after it.
2. Check your editorial process honestly if you’re relying on the human-review exception. A light read-through before publishing does not qualify; the Commission’s guidance is explicit that superficial checks don’t meet the standard.
3. Don’t confuse the provider grace period with your own deadline. If you’re a deployer publishing content, 2 August is your date — the December extension belongs to the tool vendors, not to you.

This is a two-day-out reminder, not a new rulebook — the full guide has the placement templates, the EU’s official icon set, and the editorial-responsibility test in detail.

Related reading

See the EU AI Act checklist for small business for how Article 50 fits alongside the other five articles that actually apply to ordinary teams.

Want more guides like this? Browse all Perspectives or get in touch →