Do AI-made posts need a label? The EU AI Act and Meta's rules
If AI makes my Instagram posts, do I have to say so?
Short version: realistic AI video, audio or images of people, places or events that could pass as real need a clear label. Illustrations, typography and obviously designed graphics are a much smaller concern. Two sets of rules apply to a business in the EU posting on Instagram or Facebook: the EU AI Act, and Meta’s own policy.
This is a plain-language summary, not legal advice. If your posts touch politics, health claims or real people, ask a lawyer.
What the EU AI Act says
The transparency rules are in Article 50 of the AI Act. According to the European Commission, they apply from 2 August 2026. Two parts matter for social media:
- Providers of AI tools that generate images, audio, video or text must mark the output “in a machine-readable format” so it can be detected as AI-generated. That is a job for the tool maker, not for you.
- Deployers (the business that uses the tool and publishes the result) that create a deep fake must “disclose that the content has been artificially generated or manipulated”. A deep fake here means content that resembles real people, objects, places or events and would falsely appear to be authentic.
The disclosure has to be “clear and distinguishable” and shown at the latest when someone first sees the content. For evidently artistic, creative, satirical or fictional work, the Commission says the disclosure can be made in a way that does not spoil the work.
There is a separate rule for AI-written text published to inform the public on matters of public interest. It does not apply when a person reviews the text and someone takes editorial responsibility for it.
What Meta asks for
Meta labels content as “AI info” when it detects industry-standard signals in the file or when the person posting discloses it. Meta says it requires people to disclose organic posts with photorealistic video or realistic-sounding audio that was digitally created or altered, and may apply penalties when they do not. Instagram and Facebook have a disclosure toggle for this when you post.
What this means for a small business
In practice:
- A realistic AI video with people in it (a customer in a shop, a pet at the vet, a street scene): label it. Both the AI Act and Meta point this way.
- Realistic AI voice or dialogue: label it.
- An AI photo that looks like a real photo of your real shop, product or team: label it, or better, use a real photo. Showing customers something that does not exist is a trust problem before it is a legal one.
- Illustrations, designed graphics, animated typography: these do not pretend to be real, so the deep fake rule is not aimed at them. You can still add a label.
- Captions written with AI help and checked by you: you are the editor. Everyday marketing captions are unlikely to count as informing the public on matters of public interest.
When in doubt, label. It costs nothing, and audiences mind being misled far more than they mind AI.
What Atonomuse does
Atonomuse cinematic story ads are made from AI-generated footage, so every one carries a visible “AI generated” mark, a line on the end card and a line in the caption, in the language of the post. You do not have to remember to add it.
Image posts and carousels use AI-generated images, and animated reels are motion design built from type and shapes. These do not carry the mark by default. Posts wait for your approval unless you switch that off, so you see each one before it goes out. If an image could pass as a real photo of your real shop, team or customers, reject it, or write to us and we will help you label it.