Understanding FTC Disclosure Requirements For AI-Generated Likenesses is now a practical necessity for brands, creators, and agencies using synthetic faces, voices, or personas in ads. In 2025, “real-looking” content can be produced at scale, raising risks of deception, misattribution, and hidden endorsements. Clear, timely disclosures protect consumers and reduce legal exposure. But what exactly must you disclose—and how should you do it?
FTC disclosure rules for AI likenesses: what triggers a required disclosure
When you use an AI-generated likeness—such as a synthetic spokesperson, cloned voice, realistic avatar, or deepfake-style reenactment—you create a risk that audiences will believe they are seeing a real person or a real experience. The Federal Trade Commission’s core concern is deception and unfairness in advertising. Disclosures become necessary when a reasonable consumer might be misled without them.
Common disclosure triggers include:
- Identity confusion: The audience could believe the likeness is a real person (or a specific real person) when it is synthetic or materially altered.
- Endorsement confusion: The content implies a real person endorses a product, appears in an ad, or provided a testimonial, when they did not.
- Material alteration: Visuals, audio, or “before/after” results have been changed in ways that affect the claim or overall impression.
- Source confusion: A chatbot, virtual assistant, or AI “agent” is presented like a human representative, especially in sales or customer service contexts.
The FTC doesn’t require special “AI language” in every case. It requires that advertising be truthful and not misleading, and that any necessary qualifying information be disclosed clearly and conspicuously. If the AI element changes the meaning consumers take away, you should treat it as disclosure-relevant.
Follow-up question: “What if the content is obviously fictional?” If a reasonable viewer clearly understands it is fictional or stylized, the risk drops. But “obvious” is narrower than many marketers assume. If the ad mimics documentary footage, news, influencer content, product reviews, or user-generated formats, you should expect higher scrutiny and stronger disclosure expectations.
AI-generated endorsements and testimonials: avoid deceptive marketing claims
AI-generated likenesses often intersect with endorsements—especially when brands generate an “influencer-like” personality, recreate a celebrity lookalike voice, or fabricate customer reviews. The FTC’s endorsement principles focus on whether an endorsement is truthful, reflects the endorser’s honest experience, and discloses material connections.
Key compliance points for AI-based endorsements:
- No fabricated experiences: If the “person” did not actually use the product, don’t present claims as first-person experience. An AI avatar cannot “personally” achieve results.
- No implied real-person approval: If a synthetic likeness resembles a real individual (even without naming them), avoid implying sponsorship or participation without authorization.
- Substantiate performance claims: “It improved my skin in two days” or “I earned $10,000 this month” requires reliable evidence. AI doesn’t lower the substantiation standard.
- Disclose material connections: If content functions like an endorsement (paid, free product, affiliate relationship, brand control), disclose that relationship in a way people notice.
Follow-up question: “Can we use AI to reenact a real customer’s story?” Yes, but do it carefully. Obtain permission from the real person, keep the reenactment accurate, and disclose that the depiction is a reenactment or AI-generated representation. If the story is composite or dramatized, disclose that too. The disclosure should prevent the audience from believing they are watching real footage or hearing the person’s real voice.
Follow-up question: “What about AI-generated ‘reviews’ on our site?” If they read like consumer reviews, they can be misleading. Use AI for summarization or moderation, not for inventing customer experiences. If you publish AI-written “testimonials,” you risk creating deceptive social proof.
Clear and conspicuous disclosure language: placement, timing, and format
Effective disclosure isn’t just about what you say—it’s about whether people will notice, understand, and remember it before they act. “Clear and conspicuous” means the disclosure is hard to miss and easy to grasp on the device and platform people actually use.
Best practices that map to FTC expectations:
- Put disclosures near the claim: If an AI avatar states a performance claim, place the disclosure on-screen at that moment, not only at the end.
- Use plain language: Avoid vague terms like “digitally enhanced” if the key issue is that the person is synthetic or the voice is cloned.
- Use persistent disclosures for video: For longer videos, repeat the disclosure and consider a persistent on-screen label.
- Optimize for mobile: Use readable font size, sufficient contrast, and enough on-screen time to be understood.
- Don’t bury it: Avoid placing disclosures behind “more,” in a footer, or in a fast-scrolling caption if the main impression is created before the disclosure appears.
Examples of disclosure language (choose what matches the risk):
- Synthetic spokesperson: “This is an AI-generated character, not a real person.”
- Voice clone: “Voice is AI-generated.” or “Audio is a simulated voice.”
- Reenactment of a real story: “AI-generated reenactment of a real customer story.”
- Composite/dramatization: “Dramatization. Scenes and dialogue are simulated.”
- Influencer-style ad: “Paid ad. AI-generated creator.” plus any required connection disclosures (e.g., affiliate).
Follow-up question: “Is ‘#AI’ enough?” Usually not. Hashtags can be unclear, and they don’t explain what is AI-generated (the person, the voice, the scene, the results). Use a direct statement that addresses the potential misconception.
Deepfakes and synthetic media compliance: consent, substantiation, and risk controls
Deepfakes and high-fidelity synthetic media can create a strong impression of authenticity. The compliance risk rises when content looks like evidence: a person “saying” something, a demonstration “proving” a feature, or a before/after “showing” results. You can reduce risk by building a control framework that covers consent, substantiation, and review.
1) Consent and rights management
- Get written permission from anyone whose likeness is used, recreated, or closely mimicked.
- Avoid lookalike confusion when the intention is to evoke a specific individual. If the resemblance drives attention, treat it as high-risk and consult counsel.
- Document licenses for datasets, stock assets, and voice models. Keep records that show you had the right to use inputs and outputs.
2) Claim substantiation and evidence
- Substantiate before publishing: If the ad implies objective outcomes, you need reliable support, not just “the model said so.”
- Separate demonstration from simulation: If you simulate product performance (e.g., CGI “results”), label it so viewers don’t treat it as a real test.
- Be careful with health, finance, and safety: These categories often require stronger evidence and invite higher scrutiny.
3) Review and approval workflow
- Pre-flight check: Verify disclosures appear on every placement (paid social, organic posts, CTV, affiliate sites).
- Platform compatibility: Ensure disclosures remain visible after cropping, auto-captioning, or reposting.
- Version control: Store the final ad file and screenshot the disclosure appearance as proof.
Follow-up question: “If we disclose it’s AI, can we say anything we want?” No. Disclosure does not cure false or unsubstantiated claims. It also doesn’t fix misleading net impressions. Disclosures support truthfulness; they do not replace it.
Influencer marketing and AI avatars: material connection disclosures that work
AI avatars are increasingly used as “always-on” brand ambassadors. If the avatar posts content that looks like creator content—reviews, unboxings, routines, recommendations—audiences may treat it like an endorsement. That puts the focus on material connections and transparency.
What to disclose in AI-avatar influencer campaigns:
- That the character is AI-generated when viewers could believe it is a real person.
- That the content is sponsored when the brand pays for placement or controls the message.
- Affiliate relationships when links generate commissions or the avatar “earns” on sales.
- Free products or perks when they could affect credibility, even if the “creator” is synthetic and controlled by a company.
Placement guidance for common formats:
- Short-form video: On-screen disclosure at the beginning plus in the caption. If the video includes claims, keep the disclosure visible near those claims.
- Live or Stories-style content: Repeat disclosures periodically; don’t rely on a single early frame.
- Static posts: Put the disclosure in the first lines of the caption so it’s visible without expanding.
Follow-up question: “If the brand owns the avatar, is it still an endorsement?” It can be. If the format mimics third-party recommendations, the content can carry endorsement-like weight. Treat it as advertising and disclose sponsorship clearly to prevent a misleading impression of independent opinion.
Building an FTC disclosure policy for AI content: documentation, training, and audits
A practical policy turns compliance into repeatable steps. In 2025, teams move fast across agencies, freelancers, and multiple platforms, so your policy should be simple, measurable, and backed by documentation.
Elements of a strong internal AI disclosure policy:
- Scope definition: List what counts as AI-generated or materially altered content (faces, voices, scripts, reenactments, background scenes, “proof” imagery).
- Disclosure decision tree: Require disclosure when AI could change the identity, authenticity, endorsement, or evidence impression.
- Approved disclosure library: Pre-approved language for different scenarios (synthetic spokesperson, reenactment, simulated demo, voice clone).
- Substantiation checklist: Evidence requirements for objective claims, results claims, and comparisons.
- Recordkeeping: Keep scripts, prompts (when relevant), release forms, substantiation files, and final creatives with disclosure screenshots.
- Training: Train marketing, social, PR, and customer support to recognize when AI triggers disclosure and when it doesn’t.
- Audits: Periodically review live ads, influencer posts, and affiliate pages to confirm disclosures remain visible and accurate.
Follow-up question: “Who should own this internally?” Assign a single accountable owner—often legal/compliance partnered with marketing ops—who can approve templates, enforce workflows, and manage incident response when something ships without proper disclosure.
FAQs about FTC disclosures for AI-generated likenesses
Do I always need to disclose that an ad uses AI?
Not always. Disclose when AI-generated or altered elements would mislead a reasonable consumer about who is speaking, whether footage is real, whether an endorsement is genuine, or whether a demonstration reflects real-world performance.
What is the safest wording for an AI spokesperson?
Use direct language that addresses the likely misconception: “This is an AI-generated character, not a real person.” If the voice is synthetic, add “Voice is AI-generated.”
If we have permission from a real person, do we still need a disclosure?
Often yes. Consent addresses rights; it doesn’t automatically prevent deception. If the ad could make viewers believe they are seeing real footage or hearing the real person’s voice when it’s actually AI, disclose the AI reenactment or simulated audio.
Can we use AI to recreate a celebrity if we don’t name them?
This is high-risk. Even without a name, a recognizable likeness can imply endorsement or sponsorship and can mislead consumers. Get legal advice, secure clear rights, and use prominent disclosures if you proceed.
Where should disclosures appear in short videos?
Put them on-screen early and near the relevant claim. Also include them in the caption. Don’t rely on end cards, tiny text, or “more” sections that many users never open.
Does a disclosure fix misleading performance claims?
No. Disclosures do not replace substantiation. You still must have reliable support for objective claims and ensure the overall net impression is truthful.
What about AI chatbots that interact with customers?
If people could reasonably think they’re speaking to a human, disclose that they’re interacting with AI. This is especially important when the chatbot handles sales, pricing, subscriptions, or complaint resolution.
FTC compliance for synthetic media comes down to one rule: don’t let AI change the ad’s meaning without telling people. Use disclosures when a likeness, voice, or scene could be mistaken for real, and pair transparency with strong claim substantiation. In 2025, teams that document rights, standardize disclosure language, and audit placements reduce risk and build credibility—before regulators or audiences force the issue.