Lesson 5 of 6
Disclosure & the law
6 min read
You made a slick AI ad with a synthetic presenter. Do you have to tell people it's AI? Increasingly, yes — and the law is starting to spell out exactly when.
Rules set by risk
The [EU AI Act](glossary://eu-ai-act) is the biggest attempt so far to regulate AI, and its core idea is simple: judge the use, not the tech, by how much harm it could do. A few uses are banned outright; some high-stakes ones face strict checks; many just have to be transparent — and most everyday tools are left alone.
The Act sorts uses by risk. AI media usually falls in the tier that says: you must disclose it.
Label your AI media
Under Article 50 — its transparency rules land around August 2026 — you generally must clearly label deepfakes and AI-generated or heavily AI-edited media, and tell people when they're talking to an AI. Rules differ by country and are still arriving — but "just disclose it" is fast becoming the safe, expected default everywhere.
When in doubt, disclose. Labelling AI media is quickly becoming the baseline expectation, law or not.
If you publish AI-made or AI-edited media — especially anything with a real-looking person — label it clearly. It's increasingly the law, and always the trustworthy move. (General explanation, not legal advice.)
The gist
- —The EU AI Act regulates AI by risk tier — banned, high-risk, transparency, minimal.
- —AI media usually lands in the transparency tier: you must disclose it (Art. 50, ~Aug 2026).
- —Rules vary by country, but "disclose it" is becoming the default everywhere.
You post a realistic AI-generated video of a person. Under rules like the EU AI Act, what's expected?
Continue in the app
Take the whole AI Media: Ownership, Deepfakes & Trust course — tracked
Get your personalized path, progress and streaks in the app — this lesson and every next one, in order.