Is It Legal to Make a Talking Photo? Consent and Deepfake Rules (2026)
Making a talking photo is legal in most countries when the person in the photo agrees (or it is your own face), you don't use it to deceive or harass anyone, and you label it as AI-made where the law or the platform asks you to. What turns it illegal is the use, not the tool: sexual deepfakes, using someone's likeness to sell something without consent, fake testimonials, and passing it off as real.
This guide sets out the rules that apply in 2026, each linked to the law or the platform policy itself. It is a plain-English summary, not legal advice; if you are using someone else's face in a paid campaign, talk to a lawyer where you are.
Do I need permission to make a photo of someone talk?
Yes, in practice. If the face belongs to someone else, get their consent before you animate it, and get it in writing if the video will be used commercially or published. Your own face needs nobody's permission but yours, and purely personal use, such as a birthday clip for a family group chat, is treated far more lightly than anything you publish or sell.
The legal basis: a face in a photo is personal data under the EU's GDPR (Article 4(1)), and in the US, using a person's likeness to advertise or sell something without consent is a civil wrong under right-of-publicity laws such as California Civil Code §3344, which sets damages of at least $750 per use.
Purely personal or household use falls outside the GDPR altogether (Article 2(2)(c)).
TalkPix's Terms make you responsible for holding the rights and consent for every photo and voice you upload, and scripts with sexual, abusive, hateful or fraudulent content are refused before anything renders.
Do I have to label an AI talking photo as AI?
Increasingly, yes. In the EU, anyone who publishes a realistic AI video of a real-looking person must disclose that it is AI-generated, and TikTok, YouTube, Instagram and Facebook all ask creators to label realistic AI content of real people, wherever they post from.
In the EU, Article 50 of the AI Act has applied since 2 August 2026: anyone who publishes a deepfake — realistic AI-generated or altered video of a real-looking person — "shall disclose that the content has been artificially generated or manipulated" (AI Act Service Desk, Article 50; European Commission FAQ). For evidently artistic, satirical or fictional work, a light-touch disclosure that does not spoil the work is enough.
The big platforms ask the same thing everywhere:
| Platform | What you must do | Source |
|---|---|---|
| TikTok | Label content that uses AI to show realistic-looking people or scenes; unlabeled content may be removed or labeled for you | Community Guidelines |
| YouTube | Disclose realistic content that makes a real person appear to say or do something they didn't | YouTube Help |
| Instagram and Facebook | Use Meta's AI disclosure tool for photorealistic video or realistic audio; penalties may apply if you don't | Meta newsroom, Feb 2026 |
A talking photo of a real person, posted as if it were real footage, is exactly what these rules are about. Turn the AI label on when you post.
Can I make a talking photo of a celebrity?
Not for anything that sells or endorses, and not without being obvious that it is AI. Using a famous person's face or voice to promote something without their permission is exactly what right-of-publicity and endorsement rules target, and several US states now protect a person's voice as well as their image.
A celebrity's likeness is the textbook case for right-of-publicity claims, and the US Federal Trade Commission says that using a celebrity's AI avatar without permission violates its testimonial rule if viewers would believe the celebrity actually gave the endorsement (FTC Q&A on the Consumer Reviews and Testimonials Rule).
Tennessee's ELVIS Act, in force since 1 July 2024, goes further and protects a person's voice — including a simulation of it — as property (bill text; Governor's announcement). A federal version, the NO FAKES Act, cleared the Senate Judiciary Committee in June 2026 but is not law (bill status).
Satire and obvious parody sit in a different place legally, but where that line falls is a judgment call for a lawyer, not a tool.
Can I use an AI avatar in an ad?
Yes, as long as what the avatar says is true. US advertising rules allow AI presenters and avatars in marketing; what they ban is a fake review or testimonial, including one put in the mouth of a person who does not exist.
The FTC says plainly that its rule "has no blanket prohibition on the use of AI-generated avatars in marketing" (FTC Q&A). What is banned, since 21 October 2024, is a fake review or testimonial — including one from a person who does not exist — and the Endorsement Guides require that an endorsement reflect the honest experience of the endorser (Endorsement Guides, 2023).
So an AI presenter reading your product's real features is fine. An AI "customer" describing results nobody had is not. If you turn a real written review into a video, present it as a reading of that review, not as footage the customer filmed.
Is it legal to make a talking photo of someone who has died?
Usually, for private and family use. A memorial video of a grandparent is not what these laws target, though it is kind to ask the rest of the family first. Commercial use of a well-known person's likeness after death is restricted for decades in some US states.
The GDPR does not apply to the personal data of deceased persons (Recital 27), though some EU countries add their own rules.
Commercial use is different. California protects the likeness of a "deceased personality" — someone whose identity had commercial value — for 70 years after death, and since 2025 AB 1836 adds liability of at least $10,000 for an unauthorized digital replica of one in film, TV, games or recordings (Senate Judiciary analysis; Governor's announcement). Ask the family before you publish, and keep it respectful.
What kinds of talking photo are illegal?
Sexual deepfakes, almost everywhere. Publishing intimate AI images of a real person without consent is a federal crime in the US, and in England and Wales even creating one is an offence. Fraud, defamation, harassment and impersonation laws also apply to AI video exactly as they do to anything else.
In the US, the TAKE IT DOWN Act (signed 19 May 2025) makes it a federal crime to knowingly publish non-consensual intimate images of an identifiable person, including AI "digital forgeries", and requires platforms to remove reported images within 48 hours (Public Law 119-12). In England and Wales, creating a sexually explicit deepfake of an adult without consent has been a criminal offence since 6 February 2026 (Data (Use and Access) Act 2025, s.138; commencement), and sharing one has been since January 2024 (Online Safety Act 2023, s.188).
A talking photo of your boss "announcing" something they never said can be all four of those at once.
A simple checklist before you post
- Whose face is it? Yours, or someone who said yes. Get it in writing for anything commercial.
- Whose voice is it? Your own, a built-in AI voice, or a voice you have permission to clone.
- Is anything in it false? No fake testimonials, no claims the person didn't make.
- Could a viewer think it's real footage? Turn on the platform's AI label, and in the EU, disclose it.
- Is it sexual, harassing or impersonating? Then don't make it — it is illegal in most places and against TalkPix's Terms.
For the difference between a consented talking photo and a deepfake, see talking photo vs. deepfake.
Animate a photo you have the rights to
Upload your own photo — or one you have permission to use — and make it talk in your voice or an AI voice. One-time credit packs, no subscription required.
Make a talking photo
