Is Using Someone’s Face Or Voice Without Permission Ever Legal?

Sometimes. But there is no single rule that makes an AI-generated video automatically legal or illegal.

The answer can depend on why the video is made, where it is published, how the person is portrayed, whether the face or voice is identifiable, whether the person is a public figure, and which country or state’s laws apply.

AI does not remove existing legal rights. A synthetic face, cloned voice, digital avatar, or manipulated recording can still create legal issues involving privacy, publicity rights, defamation, copyright, fraud, consumer protection, and deceptive media laws.

The key question is not simply:

“Did AI create the video?”

It is:

“What legal right allows me to use this person’s identity in this particular way?”

Which Legal Rights Can Be Triggered By An AI Face Or Voice?

A person’s face and voice can implicate different legal protections.

Legal issue	What it protects	AI-video risk
Right of publicity	Commercial use of identity	Selling or promoting something using a person’s likeness or voice
Privacy law	Certain personal or private information/interactions	Creating or distributing sensitive synthetic content
Defamation	False statements that harm reputation	Making a realistic person appear to say or do something false
Copyright	Protected creative works	Copying video, photographs, music, or other protected material
Trademark law	Brand identifiers and commercial confusion	Using a person’s identity in ways suggesting endorsement
Consumer-protection law	Against deceptive commercial practices	Fake endorsements or misleading advertising
Fraud/impersonation law	Deceptive identity use	Pretending to be the person to obtain money or access
AI/deepfake law	Certain synthetic or manipulated media	Disclosure, consent, election, or intimate-image requirements

These rights can overlap. A single AI video can create several legal problems at the same time.

Does “It Is AI-Generated” Protect Me?

No.

Calling something AI-generated does not automatically create an exception to privacy, publicity, defamation, copyright, or other laws.

Consider three examples:

Scenario	Legal risk
Clearly fictional parody using a recognizable celebrity	Potentially lower, but jurisdiction-specific
Fake video making a real person endorse a product	High commercial and deceptive-use risk
Synthetic video falsely showing a private person committing a crime	High defamation/privacy risk
Educational demonstration using a clearly labeled synthetic example	Often lower risk
Voice clone used to trick customers into sending money	Extremely high risk

The surrounding facts matter more than the fact that AI was used.

Can I Use A Celebrity’s Face Without Permission?

A celebrity’s public status does not mean their identity is free for unrestricted commercial use.

Many jurisdictions recognize some form of publicity or personality interest that can restrict unauthorized commercial exploitation of someone’s identity.

A major distinction is often:

Expression or commentary about the person versus using the person’s identity to sell something.

For example:

“Imagine what this famous actor would say about artificial intelligence.”

is legally different from:

“This actor recommends our AI software.”

The second statement can create an endorsement or commercial-use problem even if the video is entirely synthetic.

Can I Clone Someone’s Voice For A Parody?

Possibly, but parody is not a universal legal shield.

The legal analysis can depend on whether the audience understands that the work is fictional, whether the use is commercial, whether the work comments on the person, and whether the imitation creates a realistic false impression.

Use	General risk profile
Clearly fictional parody	Lower in some jurisdictions
Political satire	Potentially protected, but fact-specific
Commercial advertisement	Higher
Fake endorsement	Very high
False criminal accusation	Very high
Private-person imitation	Potentially high
Fraudulent impersonation	Extremely high

A disclaimer can help clarify the nature of a work, but a disclaimer does not automatically cure an otherwise unlawful use.

What If I Only Use The Face Or Voice For Five Seconds?

Duration does not automatically determine legality.

A short clip can still be legally significant if the identity is central to the message.

For example, a five-second synthetic voice saying:

“I personally recommend this investment.”

could be more legally problematic than a two-minute fictional parody.

The more useful questions are:

• Is the person identifiable?
• Is the use commercial?
• Does it imply endorsement?
• Is the statement false?
• Could viewers reasonably believe the person actually said it?
• Does the content involve sensitive or private subject matter?
• Does local law regulate this type of synthetic media?

What If The Video Clearly Says “AI-Generated”?

That can reduce deception risk, but it is not a universal legal defense.

Disclosure can be valuable because it helps prevent viewers from believing the person actually spoke or appeared in the video.

A stronger disclosure might be placed where viewers will actually see it rather than hidden in metadata or a caption.

Disclosure approach	Practical strength
Visible on-screen disclosure	Stronger notice
Clear description/caption	Useful
Description only on another page	Weaker
Tiny hidden text	Weak
No disclosure	Higher deception risk

The legal effect of disclosure depends on the law and facts.

Can I Use A Person’s Face For Commentary Or Criticism?

This can receive stronger legal protection in some jurisdictions, especially when the use is genuinely related to news, criticism, commentary, satire, or other expressive activity.

But the protection is not unlimited.

Courts may examine whether the person’s identity is being used primarily to communicate an expressive message or mainly to exploit the person’s identity commercially.

A useful distinction is:

Using a person’s identity to discuss the person versus using the person’s identity as a selling tool.

That distinction can be difficult in social media, where entertainment, advertising, commentary, and commerce often overlap.

What If The AI Makes The Person Appear To Say Something They Never Said?

This is one of the largest legal risks.

A realistic synthetic statement may create a false factual impression. Depending on the circumstances, that can raise questions involving defamation, false light, privacy, election laws, consumer deception, or other causes of action.

The risk becomes especially serious when the video depicts:

• Criminal conduct
• Sexual conduct
• Professional misconduct
• Political statements
• Fraud
• Medical claims
• False endorsements
• Statements damaging reputation

The technical realism of the AI can make the legal problem worse because viewers may reasonably believe the recording is authentic.

What If I Do Not Publish The Video?

Private creation is generally different from public distribution, but it does not mean every use is automatically lawful.

Legal exposure can depend on what you do with the material, why it was created, how it was obtained, and whether other laws regulate the underlying conduct.

For example, creating a voice clone for personal experimentation can present a very different risk from using it to impersonate someone in a financial transaction.

Can I Use Someone’s Voice If Their Voice Is Not Copyrighted?

This is a common misunderstanding.

Copyright generally protects qualifying works of authorship, not simply every personal characteristic.

A person’s voice itself is not necessarily treated as a copyrighted work merely because it is recognizable.

However, that does not mean a voice is legally unrestricted.

Other laws may regulate unauthorized commercial use, impersonation, deceptive conduct, privacy, or publicity.

The question is therefore not:

“Is the voice copyrighted?”

It is:

“What legal right controls the particular use of this person’s identity and voice?”

What About Using A Real Photo As The Training Source?

That creates another layer of analysis.

Using a photograph to create an AI likeness can involve rights in both:

1. The person depicted, and
2. The photograph itself.

The photographer, agency, publisher, or other rights holder may have copyright interests in the original image.

Component	Possible legal issue
Person’s face	Publicity/privacy/personality rights
Original photograph	Copyright
Voice recording	Copyright or contractual rights may exist in the recording
Generated output	Depends on jurisdiction and underlying material
AI model	License/terms may restrict use
Uploaded material	Platform or contract restrictions

This is why an AI output can create legal problems even when the final video looks “new.”

Can I Use A Person’s Identity For A Fake Advertisement?

This is one of the clearest high-risk situations.

Imagine an AI-generated actor who looks and sounds exactly like a real celebrity saying:

“I use this product every morning.”

Even if the celebrity never appeared in the recording, the video can create a false impression of endorsement.

That can potentially trigger:

• Publicity-right claims
• False endorsement claims
• Consumer-protection concerns
• Advertising-law issues
• Defamation concerns
• Contractual disputes

A realistic AI representation should therefore not be treated as a substitute for an actual endorsement agreement.

What If I Have A License From The AI Platform?

A license from an AI platform may govern your use of the software or generated output, but it does not necessarily give you permission from the person being depicted or imitated.

These are separate rights.

Permission	What it may cover
AI platform license	Use of the AI service and possibly output
Photographer license	Use of a particular image
Actor’s release	Use of the actor’s likeness/voice
Publicity license	Commercial identity use
Music license	Music recording/composition
Model release	Certain image uses

You may need permission from multiple rights holders.

What Is The Safest Way To Make An AI Video Without Permission?

The safest strategy is to avoid making the real person identifiable.

Instead, use:

A fictional face + fictional voice + fictional name + fictional biography

If you need realism, create an original synthetic person rather than copying a real individual’s identifiable characteristics.

For higher-risk projects, obtain written permission or a properly drafted license that specifically addresses:

• Face and likeness
• Voice
• AI cloning
• Synthetic media
• Advertising
• Social media
• Duration
• Geographic territory
• Platforms
• Editing rights
• Future AI model use
• Revocation or termination

What Should A Written AI Likeness Agreement Actually Cover?

A vague release such as “permission to use image” may not fully address AI.

A technically detailed agreement should identify the rights being licensed.

Contract term	Why it matters
Likeness	Defines visual identity rights
Voice	Covers voice cloning or synthesis
AI training	States whether data may train models
AI generation	States whether synthetic replicas are allowed
Commercial use	Covers advertising and monetization
Editing	Defines permitted transformations
Endorsement	Specifies whether implied endorsement is allowed
Territory	Defines geographic scope
Duration	Defines license period
Platforms	Covers websites, social media, ads, etc.
Approval rights	Determines whether the person can approve final content
Exclusivity	Prevents or permits competing uses
Revocation	Defines what happens after termination

What Technical Questions Should I Ask Before Publishing?

Before publishing a realistic AI face or voice, ask:

• Identity: Can an ordinary viewer recognize the person?
• Consent: Do I have documented permission?
• Purpose: Why am I using this identity?
• Commercial use: Am I selling, promoting, or monetizing something?
• Truth: Does the video make a factual claim about the person?
• Endorsement: Could viewers think the person approved something?
• Disclosure: Is the synthetic nature clearly disclosed?
• Source rights: Do I have rights to the photos, videos, and audio used?
• Jurisdiction: Which country’s or state’s law applies?
• Platform rules: Does the distribution platform have AI/deepfake restrictions?
• Sensitive content: Does the video involve politics, sex, crime, health, or other high-risk subjects?

Which Uses Are Usually Higher Risk?

AI use	Risk level	Main concern
Fictional person	Low	Avoids real-person identity issue
Licensed actor/avatar	Lower	Depends on license
Clearly labeled parody	Medium	Context and jurisdiction matter
News/commentary	Medium	Expression and accuracy questions
Celebrity commercial imitation	High	Publicity/endorsement
False statement about a real person	High	Defamation/privacy
Sexual deepfake	Very high	Consent, privacy, specialized laws
Political deepfake	Very high	Election and deceptive-media rules
Financial impersonation	Extremely high	Fraud/identity deception

These are risk categories, not universal legal conclusions. Local law can change the result.

Is There One Universal “AI Deepfake Law”?

No.

The legal framework can vary by country, state, province, and subject matter. Some jurisdictions have enacted specific synthetic-media rules, while others rely primarily on older laws concerning publicity, privacy, defamation, fraud, copyright, or consumer protection.

That means a legal answer should always begin with:

Where will the person be located, where will the content be published, and what is the purpose of the AI video?

A video lawful in one jurisdiction may create significant exposure in another.

What Is The Simple Rule For Using Someone’s Face Or Voice?

Use this rule:

The more realistic, commercial, deceptive, sensitive, or damaging the AI representation is, the more important consent and legal review become.

The safest path is not to ask:

“Can AI technically make this?”

It is to ask:

“Do I have the legal right to make this particular person appear or sound this way?”

AI can make a synthetic identity look real in seconds. Legal permission is a separate problem.

FAQs

Can I legally clone a celebrity’s voice?

Sometimes, depending on the jurisdiction and use. Commercial, deceptive, or endorsement-related voice cloning creates substantially higher risk than some forms of commentary or parody.

Can I use a private person’s face in an AI video without permission?

That can create serious privacy, publicity, defamation, or other legal issues, especially when the person is identifiable and the content is misleading, commercial, sexual, or damaging.

Does adding “This is AI” make the video legal?

No. Disclosure can reduce deception, but it does not automatically eliminate publicity, privacy, defamation, copyright, or other legal claims.

Can parody protect an AI deepfake?

Potentially, but parody is highly fact-specific. A court may consider whether the audience would understand the work as parody or reasonably believe it is an authentic statement.

Can I use someone’s face in an AI-generated news video?

Potentially, especially for genuine commentary or reporting, but the exact legal protections vary by jurisdiction and the content must still avoid unlawful conduct such as certain defamatory or deceptive uses.

Can I use someone’s voice if I never use their name?

Maybe not. A person can be identifiable from a distinctive voice, face, mannerisms, or combination of characteristics even without an explicit name.

Is a public figure’s image free to use?

No. Public visibility does not automatically eliminate publicity, privacy, trademark, or other legal protections.

Who owns an AI-generated person’s likeness?

That question can involve several different rights. The AI output, the underlying source material, and the real person’s identity can be controlled by different legal interests.

Do I need permission to use someone’s old YouTube video to train or create an AI imitation?

Possibly. You may need to consider copyright, platform terms, privacy, publicity rights, contractual restrictions, and the laws governing the resulting use.

What is the safest option?

Use a completely fictional face and voice, use properly licensed actors or synthetic characters, or obtain written consent that expressly covers AI-generated likeness and voice use.

When should I talk to a lawyer?

Get jurisdiction-specific legal advice before publishing high-risk material involving a real person’s identity, especially commercial endorsements, political content, sexual material, criminal accusations, financial impersonation, or content likely to damage reputation.