Free · No signup · 2-minute check

The EU AI Act delay? It probably doesn't cover you.

Headlines said the AI Act was postponed — but that was the high-risk rules. The Article 50 transparency duties — chatbot disclosure, AI-content labels, deepfake notices — still apply from 2 August 2026, including to small businesses. Check what hits you, then generate your disclosure pack.

days
hours
mins
Aug 2
deadline

Was your obligation delayed? Answer honestly:

Toggle everything your business does. The verdict updates instantly — with the real dates.

We have a chatbot or AI assistant that talks to customers
Website chat widget, AI phone/voice agent, AI support assistant — including third-party ones you deploy
We publish AI-generated images, video or audio
Product visuals, social posts, ads, voiceovers, music — generated or heavily manipulated by AI
We publish AI-written text on news or public-interest topics
Blog/news articles informing the public — without full human editorial review
We create realistic AI depictions of real people, places or events
Face swaps, cloned voices, "photo-real" scenes that could be mistaken for authentic
We use emotion recognition or biometric categorisation
AI reading customer emotions, or sorting people by biometric traits

Generate your disclosure pack

Ready-to-paste text for every duty you toggled above. Edit the details — the pack builds live.

Your business
Details (only used where relevant)
Your disclosure pack
Toggle what you do in the checker above, then add your business name — your disclosure pack appears here.

What actually happened with the "delay"

In 2026 the EU agreed to postpone the AI Act's high-risk system obligations (to Dec 2027 / Aug 2028). Headlines compressed that into "the AI Act is delayed." But Chapter IV — Article 50 transparency — was not part of that postponement. From 2 August 2026, people must be told when they're talking to AI, and AI-generated content must be disclosed. The single carve-out: machine-readable marking of synthetic content, for generative systems already on the market before 2 Aug 2026, got until 2 December 2026. Everything else lands on time.

💬
Chatbots must say they're AI

People interacting with an AI system must be informed they're talking to AI, unless it's already obvious from context.

🏷️
AI content must be disclosed

Deepfakes and AI-generated/manipulated media need clear disclosure; AI text on public-interest matters needs a notice unless there's human editorial review.

📡
Marking = the only delay

Machine-readable marking of synthetic output got a Dec 2, 2026 extension — only for systems on the market before Aug 2. Your disclosure duties still start Aug 2.

I'm a small business using someone else's chatbot. Do I care?

Yes. The transparency framework covers deployers, not just AI companies. If your customers chat with an AI on your website, the safe position is a clear "you're talking to an AI" notice at first interaction — which is exactly what this pack generates. Fines under the AI Act can reach €15M or 3% of worldwide turnover, and enforcement starts with exactly this kind of visible, checkable duty.

Does a human reviewing our AI blog posts change anything?

Yes — AI-assisted text published to inform the public on matters of public interest is exempt from the disclosure where the content has undergone human review and someone holds editorial responsibility. If nobody reviews it before publishing, disclose it.

What's the December 2, 2026 date about?

That's the extended deadline for the machine-readable marking requirement (watermarks/metadata in synthetic output) — and only for generative systems placed on the market before 2 August 2026. It is not a delay of your disclosure duties.

Is this legal advice?

No — see the note below. It's a structured starting point built from the Regulation's text and reputable legal analyses, designed so nothing obvious is missing when you paste disclosures into your site.

Not legal advice. This free tool summarises Article 50 transparency duties and generates disclosure text templates. It does not assess your specific systems, does not cover high-risk obligations, and does not guarantee compliance with the AI Act or national implementations. Verify current requirements with counsel before relying on generated text. Sources: Regulation (EU) 2024/1689, Art. 50 & Ch. IV; 2026 Digital Omnibus provisional agreement.
Copied to clipboard