Since 2 August 2026, the transparency rules of the EU AI Act — Article 50 — apply to every business that puts AI in front of people. If your website has a chatbot, if your team sends AI-drafted emails, or if your systems generate documents with AI, this article concerns you.

The good news: for most businesses, honest compliance takes an afternoon, not a legal project. Here is what the rule actually asks for, who is responsible for what, and a checklist you can run in five minutes. And because we would not ask you to do anything we haven’t done ourselves, we’ll also show you exactly how the Mitigate AI Platform handles Article 50 — and what we found when we audited ourselves and the big model labs.

The short version
  • Tell people when they're talking to AI.
  • Put an invisible "made by AI" stamp on AI-generated files — think of it as a nutrition label software can read.
  • If you make deepfakes or publish AI-written news, say so. Most businesses never touch this one.

What the rule actually asks for

Article 50 boils down to three obligations. No badge on every sentence, no product rebuild — just being honest about where AI shows up.

50(1) — DISCLOSE

Tell people they're talking to AI

Right in the conversation, by the first message — unless it's already obvious they're dealing with AI.

On the provider
50(2) — MARK

Mark AI-generated content

AI-generated files carry a machine-readable marker so software can spot that they were made by AI.

On the provider
50(4) — LABEL

Label deepfakes and AI-written news

Deepfakes and AI-written public-interest articles must be labelled — unless a human reviewed the content and holds editorial responsibility.

On the deployer — rarely applies

Who’s actually responsible

The provider builds and ships the AI system. For our clients, that’s Mitigate for the platform, and the model labs (Anthropic, OpenAI, Google) for the engine underneath. Telling users it’s AI and marking the output — that’s the provider’s job. We carry it, so you don’t have to build any of it.

The deployer is you, using the AI in your business. Your one extra duty only kicks in if you use AI to make deepfakes or publish articles on public-interest topics — then you label them. For everyday business use, it doesn’t come up.

The person chatting is the one being protected. They carry no obligations at all.

A quick myth-buster: “the model is Anthropic’s, so it’s their problem” doesn’t hold — as a deployer, you still have your part. And if you white-label or heavily rework an AI system under your own brand, you may step into the provider’s shoes yourself.

Your 5-minute checklist

  • List the AI that faces people. Your website chatbot, AI-drafted emails, AI-generated reports or offers. For each, check two things: does it tell people it's AI, and is its output marked as AI-made? Every "no" is something to fix.
  • Make sure the AI notice is real. Visible in the chat, not buried in the fine print. A vague "assistant" or a hidden tag doesn't count.
  • Check it's accessible. Can a screen-reader user hear it? The rule requires it.
  • Ask your vendors — in writing. Do they mark AI-generated content? Keep the reply on file.
  • Flag AI-written public-interest text, if you publish any.
  • Don't over-label. Internal, obviously-AI tools don't need a badge on everything.
  • Watch the clock. The AI notice (50(1)) and deepfake labelling (50(4)) apply from 2 August 2026, no delay. Systems already on the market get a grace period for the machine-readable marking.

We ran this on ourselves — and on the big guys

Before writing this, we audited our own platform against the checklist above. Two findings are worth sharing.

Our chatbot already showed the AI notice. It had been sitting under the chat all along — we just needed to make it editable and translatable per language. Lesson: check what you already have before building anything.

The most popular AI tools don’t mark generated documents at all. We asked ChatGPT and Claude to generate a Word file, then opened its properties. Both were blank — no “made by AI” marker anywhere. So the moment you do even a little, you’re ahead of the pack.

How the Mitigate AI Platform handles it for you

If you run your assistants and agents on the Mitigate AI Platform, the provider-side obligations of Article 50 are already covered. Here’s what the platform does out of the box.

It tells users it’s AI. A clear notice sits under the chat, from the first message, on by default: “AI assistant can make mistakes. Check important info.” It’s editable per language so you can localise it — and switching it off is a deliberate, logged choice, not an accident.

It nudges people to double-check. The same line does double duty, the way Claude and ChatGPT do it: disclosure plus a healthy reminder that important information deserves verification.

It stamps generated files. Documents generated on the platform carry a machine-readable “Generated with AI (Mitigate AI Platform) — EU AI Act Art. 50(2)” mark in their properties, plus a visible footer for official-use files. For PDFs we use C2PA Content Credentials — the industry provenance standard, which covers PDFs and documents, not just images.

Images and audio aren’t in scope for us today — the platform doesn’t generate them, so that kind of watermarking isn’t ours to worry about. If it ever becomes relevant, we’ll add C2PA there too.

The honest part

We’d rather tell you about the limits than oversell.

File metadata is best-effort. It can be stripped, and heavy manual editing may wipe it — which is exactly why we add a visible footer as backup for official-use documents. For context: as of mid-2026, files from ChatGPT and Claude carry no marking at all, so platform-stamped files are already at or above the industry line.

Invisible text watermarking isn’t ready yet. Today it’s basically Google’s SynthID on Gemini; OpenAI and Anthropic don’t watermark text. The law only asks for what’s “technically feasible,” so we’re watching the standards and will add it the moment something actually holds.

Our rule of thumb

Do what the law asks — and nothing performative. Clear AI disclosure, machine-readable marking where it counts, and no badge-spam that clutters the experience. You stay compliant; your users stay happy.

Want your AI transparency handled by the platform instead of by hand? Get in touch. We’ll walk through the checklist with you, show you the platform’s disclosure and marking features live, and map anything that’s left to fix.

Sources and further reading

A practical summary, not legal advice. Last fact-checked 5 August 2026.