Do you legally have to say your influencer is AI? (In Europe, yes)

Most brands ask us the fun question first: does an AI influencer actually work? (Short answer: sometimes spectacularly.) The question they should ask second — and almost never do — is the boring one that can cost you €35,000: do I legally have to tell people it’s fake?

In Europe, the answer is now clearly yes. And the deadline is closer than most marketers realise.

The law caught up in August 2026

The EU AI Act sorts AI systems by risk. Virtual influencers, synthetic voices, generated video — all of it lands in the “transparency risk” bucket, which comes with a hard rule: people have the right to know when they’re looking at something a machine made.

Article 50 spells it out. Providers of generative AI must ensure outputs are “marked in a machine-readable format and detectable as artificially generated,” and deployers of deepfakes — which legally includes a fabricated human persona endorsing your product — must disclose that the content was artificially created or manipulated. These transparency obligations enter into force on 2 August 2026.

The European Commission’s own summary is blunt: certain AI-generated content, “namely deep fakes and text published with the purpose to inform the public,” must be clearly and visibly labelled. A brand ambassador who has never existed is squarely inside that definition.

So if you’re running virtual-persona campaigns to a Spanish or EU audience, “we’ll disclose it if anyone asks” is no longer a strategy. It’s a compliance gap.

This isn’t just an EU thing

If your reflex is “great, we don’t sell in Europe” — slow down. The disclosure principle is basically universal now:

• Spain has gone further than most, drafting its own national law that treats failure to label AI-generated content as a serious infringement, with fines reaching into the tens of millions of euros for the worst cases.
• The US doesn’t have an AI-influencer statute, but the FTC’s endorsement rules already ban deceptive endorsements and fake testimonials. A synthetic person “recommending” your product with no disclosure is a textbook deceptive endorsement.
• The platforms (Meta, TikTok, YouTube) all now require creators to flag realistic AI-generated content, independent of any government.

The pattern is clear: every serious jurisdiction and platform is converging on the same rule. Disclose, or you’re exposed.

The good news: disclosure barely costs you anything

Here’s the part brands panic about for no reason. Every case study where AI personas actually performed had one thing in common — the audience knew. The novelty was the appeal. Hiding it doesn’t buy you engagement; it buys you legal risk and a betrayal moment when people figure it out.

Audiences smell a fake human instantly and punish it. They’re remarkably forgiving of an openly artificial one. Transparency isn’t the tax on the campaign — it’s the thing that makes the campaign safe to scale.

How to actually comply (without ruining the vibe)

You don’t need a lawyer on retainer for this. You need four habits:

• Label the account, not just the posts. Bio line: “Virtual creator / AI persona.” Persistent, unmissable, done once.
• Use the platform’s AI toggle. Every major network has a “made with AI” or “AI-generated” switch. Flip it. It satisfies most of the disclosure burden automatically.
• Keep the disclosure visible, not buried. A #AI hashtag as the 30th tag in a wall of hashtags won’t cut it under a “clear and visible” standard. Put it where a normal person actually reads.
• Keep receipts. Note which tools generated what, and when. If a regulator or platform ever asks, “we can show our process” is a very different conversation from “we’ll get back to you.”

That’s it. None of this dents performance. It just moves you from exposed to bulletproof.

Where the real risk hides

The dangerous move isn’t running an obvious, labelled AI persona. It’s the quiet stuff:

• A “customer testimonial” video where the customer is synthetic and unlabelled.
• AI-generated “reviews” seeded across your product pages.
• A founder’s talking-head clip that’s actually a deepfake of a real person’s face.

Those aren’t edgy marketing — they’re exactly the deception the law was written to stop, and they’re where the fines live. If you wouldn’t be comfortable putting “made with AI” on it in 24-point type, don’t publish it.

FAQ

Does a labelled AI influencer still count as a “deepfake” under the Act? For disclosure purposes, effectively yes — a fabricated persona presented as a real endorser falls under the transparency rules. The fix is the same either way: label it clearly and you’re compliant.

We only run ads, not organic. Are we off the hook? No. If the ad creative contains AI-generated people or voices presented as real, the same disclosure logic applies — and it sits right next to your Google Ads compliance, where deceptive claims already get accounts suspended.

Is this worth the hassle for a small local business? Usually the AI-persona play isn’t your first move at all — local brands win faster with solid web design, SEO and ads. But if you do use any synthetic content, the disclosure rules apply to you exactly the same as to a global brand.

The bottom line

The EU AI Act didn’t ban AI influencers. It banned pretending they’re real — which was always the version that blew up in brands’ faces anyway. Label it, toggle the switch, keep your receipts, and the whole compliance question evaporates.

Thinking about a virtual-persona campaign and not sure where the legal lines are? Tell us what you’re planning and we’ll give you a straight answer — including “don’t do this one” if that’s the honest call.