Your virtual spokesperson has never tried your product. Under FTC rules, that fact alone can turn a polished campaign into an enforcement problem. Disclosure rules for AI generated virtual spokespeople now reach across the US and EU, and the brand, not the avatar vendor, carries the exposure. Here is what you must label, where, and who should own it.
Two different questions regulators are asking
Most teams treat “AI disclosure” as one checkbox. It is actually two separate obligations, and mixing them up is how brands end up half-compliant.
Question one: is the content synthetic? That is a transparency issue, and the EU AI Act is the main driver. Question two: is the endorsement real? That is a deception issue, and the FTC is the main driver.
A virtual spokesperson can trigger both at once. The avatar looks human (question one), and it claims a personal experience that never happened (question two). Fixing the first with a small “AI” tag does nothing for the second.
The US angle: a persona that does not exist cannot testify
The FTC’s final rule on reviews and testimonials is the clearest signal. The agency states that the rule addresses testimonials that misrepresent they come from someone who does not exist, including AI-generated fake reviews, and the rule also covers people who never actually used the product. The same release says the rule lets the agency seek civil penalties against knowing violators.
Read that against a typical virtual brand ambassador. If your avatar says “I switched to this serum three months ago and my skin changed,” you have a fabricated personal experience delivered by a non-person. A disclosure label may not cure that. The claim itself has to go.
A label can tell people the spokesperson is synthetic. It cannot make a invented first-person experience truthful. Rewrite the script, then label the avatar.
So what can a virtual spokesperson safely say? Product facts you can substantiate, brand-authored claims presented as the brand’s own, and clearly framed demonstrations. What it should avoid: testimonial language like “I tried,” “I love,” or “it worked for me” unless it sits inside an obviously fictional or character-driven frame and nothing implies real use.
Our deeper look at clear and conspicuous labels covers placement and wording. The short version: if a viewer has to hunt for it, it does not count.
The EU angle: label the deepfake at first exposure
The EU is more prescriptive. According to the European Commission’s Article 50 FAQ, deployers must disclose deepfake content to a person upon first exposure at the latest. The transparency rules apply from 2 August 2026.
What counts as a deepfake? The Commission describes three cumulative criteria, including a high level of resemblance to a real person, object, place or event, and a false appearance of authenticity. Law firm Paul Weiss, summarizing the final guidelines, notes that the test is objective: intent to deceive is not required and photorealism is not decisive.
That matters for brands. An avatar that is a wholly invented character may sit outside the deepfake definition. An avatar modeled on a real creator, a celebrity lookalike, or a cloned voice almost certainly does not. If you licensed a real person’s likeness and then animated it, assume you are in deepfake territory.
What “perceivable” means in practice
The guidance says information must be perceivable in practice, not buried in terms or menus. Paul Weiss also reports that a single up-front notice may not be enough for longer formats, and the guidelines point to examples such as the start of a deepfake video. For a 60-second reel, a label in the first frame is the floor. For a long-form video or a podcast-style ad, plan for repeated cues.
There is also a carve-out. For evidently artistic, creative, satirical or fictional works, the Commission says the duty narrows to disclosing the deepfake in a way that does not hamper enjoyment of the work. Do not lean on this for a product demo. A sales pitch is rarely “evidently fictional.”
What brands must label: a working checklist
Here is the practical split I use when auditing virtual spokesperson programs.
- Synthetic face or body: Label at first exposure, on every placement, including cutdowns and paid variants.
- Cloned or synthetic voice: Label audio-only formats too. A voice-over with no visual gives viewers no other cue.
- Likeness of a real person: Treat as deepfake by default. Confirm written consent and scope. See our piece on AI likeness and voice clauses.
- Material connection: The brand owns the avatar, so say it is the brand’s own ambassador or an ad. Paid placement through a human creator still needs the standard sponsorship disclosure.
- Experience claims: Remove or reframe any statement implying real use by a non-existent person.
- Comments and replies: If an AI persona responds to customers or recommends products, identify it as AI at the start of the interaction.
That last point links to a growing issue. Automated agents that recommend products create their own disclosure questions, which we cover in AI shopping agents and material connection.
Who actually gets fined?
Not the avatar studio, in most cases. The EU framework splits duties between providers (the tools that generate content, which must mark outputs in a machine-readable format) and deployers (you, publishing it). Brands and agencies running campaigns are usually deployers.
The stakes are real. The stakes are real: according to the Commission’s Article 50 FAQ, fines can reach up to 15 million euros or 3% of total worldwide turnover for the preceding financial year, and proportionality can be taken into account for SMEs and small mid-cap companies. Enforcement sits mainly with national market surveillance authorities, so expect uneven attention across member states.
One timing wrinkle. Paul Weiss reports a deferral to 2 December 2026 for marking and detection obligations covering generative systems already on the market. That deferral belongs to providers. It does not give deployers a pass on labeling deepfakes. Content published before 2 August 2026 does not need retroactive labels, per the same guidance, but anything you re-run or re-edit after that date is a fresh publication.
Deployers are also expected to build proportionate internal processes, plus awareness and review mechanisms for their labeling. In plain terms: regulators want to see a workflow, not a one-off fix.
Build the workflow before the campaign, not after
I have watched teams scramble to add labels after legal flagged a live asset. It is expensive and it leaves a paper trail of the wrong kind. A cleaner approach has four steps.
- Classify every avatar. Fully fictional, real-person-based, or cloned voice. Each class gets a default label rule.
- Script review for experience claims. Legal or compliance signs off on any first-person language before production.
- Contract the vendor. Require disclosure of tools used, watermarking support, and likeness consents. Our guide on supplier diligence shows what to ask.
- Check insurance. Many policies carve out AI risks. Review AI exclusion clauses before you assume coverage.
If you run campaigns across markets, align on the strictest rule rather than juggling versions. We explain that logic in one disclosure standard for every market.
Design for the strictest regime you operate in. One label standard is cheaper than ten market-specific edits.
A note on platform tools
Platform toggles help, but they are not a legal shield. Platform AI labels and sponsorship toggles vary and change often, and relying on them alone leaves gaps if a feature fails or a cutdown is reposted elsewhere. Burn the disclosure into the asset itself and treat platform labels as a backup layer.
FAQs
Do I have to label a fully fictional AI avatar?
Under EU rules, a wholly invented character may fall outside the deepfake definition, but you should not assume that. In the US, the bigger risk is deception: viewers could believe a real person is endorsing the product. Labeling the avatar as AI and the post as an ad is the safest default.
Can a virtual spokesperson give a testimonial?
Not a first-person one that implies real use. The FTC’s rule targets testimonials that misrepresent they come from someone who does not exist or who never used the product. Keep claims brand-authored and substantiated.
When do the EU AI Act labeling rules apply?
The Article 50 transparency obligations apply from 2 August 2026. A limited deferral exists for provider marking and detection duties on systems already on the market, but deployer labeling of deepfakes is not covered by that relief.
Where should the label appear?
It must be perceivable, not hidden in menus or terms. For deepfake video, the guidance points to disclosure at first exposure, such as the start of the video. Longer formats may need repeated cues.
Who is liable, the brand or the AI vendor?
Typically the brand as deployer carries the labeling duty, while the vendor carries provider duties such as machine-readable marking. Contract for both and keep records of your own review process.
What are the penalties?
In the EU, penalties scale with company size, and in the US, the FTC’s reviews rule allows civil penalties against knowing violators. In the US, the FTC’s reviews rule allows civil penalties against knowing violators.
Next step: pull every live asset featuring a synthetic face or voice this week, tag each as fictional, real-person-based, or cloned, and strip any first-person experience claims before you touch the labels.
Top Influencer Marketing Agencies
The leading agencies shaping influencer marketing in 2026
Agencies ranked by campaign performance, client diversity, platform expertise, proven ROI, industry recognition, and client satisfaction. Assessed through verified case studies, reviews, and industry consultations.
Moburst
-
2
The Shelf
Boutique Beauty & Lifestyle Influencer AgencyA data-driven boutique agency specializing exclusively in beauty, wellness, and lifestyle influencer campaigns on Instagram and TikTok. Best for brands already focused on the beauty/personal care space that need curated, aesthetic-driven content.Clients: Pepsi, The Honest Company, Hims, Elf Cosmetics, Pure LeafVisit The Shelf → -
3
Audiencly
Niche Gaming & Esports Influencer AgencyA specialized agency focused exclusively on gaming and esports creators on YouTube, Twitch, and TikTok. Ideal if your campaign is 100% gaming-focused — from game launches to hardware and esports events.Clients: Epic Games, NordVPN, Ubisoft, Wargaming, Tencent GamesVisit Audiencly → -
4
Viral Nation
Global Influencer Marketing & Talent AgencyA dual talent management and marketing agency with proprietary brand safety tools and a global creator network spanning nano-influencers to celebrities across all major platforms.Clients: Meta, Activision Blizzard, Energizer, Aston Martin, WalmartVisit Viral Nation → -
5
The Influencer Marketing Factory
TikTok, Instagram & YouTube CampaignsA full-service agency with strong TikTok expertise, offering end-to-end campaign management from influencer discovery through performance reporting with a focus on platform-native content.Clients: Google, Snapchat, Universal Music, Bumble, YelpVisit TIMF → -
6
NeoReach
Enterprise Analytics & Influencer CampaignsAn enterprise-focused agency combining managed campaigns with a powerful self-service data platform for influencer search, audience analytics, and attribution modeling.Clients: Amazon, Airbnb, Netflix, Honda, The New York TimesVisit NeoReach → -
7
Ubiquitous
Creator-First Marketing PlatformA tech-driven platform combining self-service tools with managed campaign options, emphasizing speed and scalability for brands managing multiple influencer relationships.Clients: Lyft, Disney, Target, American Eagle, NetflixVisit Ubiquitous → -
8
Obviously
Scalable Enterprise Influencer CampaignsA tech-enabled agency built for high-volume campaigns, coordinating hundreds of creators simultaneously with end-to-end logistics, content rights management, and product seeding.Clients: Google, Ulta Beauty, Converse, AmazonVisit Obviously →