AI Likeness Rights in the UK Guard Your Face. They Won't Let You Sell It.
British law protects a well-known face from being faked, hijacked or degraded, but it stops short of making that face something you can own and license. Denmark is about to cross that line, and any business built on named talent should know which side of it their contracts sit on.
Ask a plain question about any business built on a named face, whether a football pundit or a founder whose photograph sells the brand: if someone manufactures that face tomorrow and puts it to work, what exactly do you own? In the UK the honest answer is uncomfortable. You own a set of remedies against being harmed. You do not own the face. AI likeness rights in the UK get discussed as if they were a settled property right waiting to be exercised. They are not. They are a patchwork of doctrines built for other purposes and pressed into this one.
That distinction, between a remedy against harm and an asset you can hold, is the whole commercial story. It decides whether a face sits on your balance sheet as something you can license or in your legal budget as something you defend. Contracts written before cheap synthesis never had to choose, because a convincing fake was expensive and rare. It is now neither, and the choice will not wait.
What do AI likeness rights in the UK actually protect?
Three separate bodies of law do the work, and none was designed for it. The first is passing off. In the Court of Appeal's decision in Irvine v Talksport, a well-known individual was held able to sue when their image is used to imply an endorsement they never gave. Eddie Irvine, the Formula One driver, won because a radio station's doctored photograph misrepresented that he backed it. The requirements are demanding: established goodwill, a misrepresentation, and resulting damage. Passing off guards a reputation that is already commercial. It does little for a private individual, and it hands you no title you can assign or license to anyone else.
The second is data protection. A face used to identify a specific person is biometric data, and the ICO's guidance on biometric data treats it as special category personal data under UK GDPR, with the stricter conditions that classification carries. This is genuine control: it governs how a face may be processed, stored and reused. It is control over processing, though, not ownership of the likeness, and it turns on identification rather than commercial value.
The third is criminal law. The Online Safety Act 2023 created offences for sharing, or threatening to share, intimate images without consent, and Parliament drafted them to cover images generated by computer, deepfakes included. That closed a real gap in how the law protects dignity. It also covers a narrow band of harm and confers no commercial right at all.
Stack the three together and the shape is clear. British law is generous with protection against specific harms and quiet on the question of ownership. You can stop a false endorsement, restrict the processing of your biometric data, and see the worst abuses prosecuted. You cannot point to a statute that says the face is yours to license.
Why Denmark's bill is the one to watch
Denmark has proposed the opposite starting point. Its government has set out plans to amend national copyright law so that people hold rights over their own facial features, body and voice, giving them a direct basis to demand the removal of realistic deepfakes and, just as important, to license legitimate use. Copyright is property. It can be owned, assigned and sold. If the bill passes close to its announced form, a Danish citizen would hold in their likeness the kind of asset a UK citizen has to assemble by hand.
That contrast is the argument. Denmark is moving to convert protection into property, while the UK offers protection without it. For a business that depends on named talent, the gap is concrete. It changes what you can write into a contract, what you can defend in court, and whether a face reads as an asset or a running cost.
What should UK businesses do before the law catches up?
The practical answer is to build the missing right privately. Since no UK statute grants an ownable likeness, contracts have to supply the equivalent: explicit consent to defined synthetic uses, a licence scope that sets out what may and may not be generated, revocation terms, and a revenue share wherever a face is used commercially. Agreements signed before synthesis went cheap say nothing about any of this, because they assumed a face could not be faked at scale. That silence is the exposure.
This decision deserves proper thought rather than a hurried clause. How consent, licensing and provenance get recorded and enforced is a matter of technical strategy as much as legal drafting, and it warrants the discipline you would bring to any AI deployment kept under human control. Consent and provenance tracking should sit on systems that are secure by construction, because a likeness licence is worth no more than the audit trail behind it.
There is a constructive side to this. The same tools that let a bad actor fake a face let a willing person open a sanctioned, revenue-shared channel for licensed synthetic content. Handled well, consent plus monetisation turns a defensive posture into a product line. The businesses that come out ahead will treat their faces as something to govern on purpose, not something to chase down after the fact.
What would change this analysis is UK legislation that follows the Danish route and grants a real, assignable likeness right. That is plausible, and worth watching. I would not build a strategy on its arrival. The base case is a UK that keeps offering remedies against harm while withholding a property right, and businesses treating a valuable face as if it were already owned may discover, in a dispute, that they hold less than they thought.
Questions people ask
Can I stop someone using an AI version of my face in the UK?
Yes in specific situations, but through separate laws rather than one likeness right. If the fake implies you endorse a product, you may have a passing-off claim, as Irvine v Talksport established, provided you have commercial goodwill. If it is an intimate image, the Online Safety Act 2023 created criminal offences that reach deepfakes. If your face is being used to identify you, UK GDPR treats it as biometric data. What you cannot yet do is claim you simply own the likeness.
Is a person's face protected by copyright in the UK?
No. UK copyright protects a particular photograph or recording, usually owned by whoever created it, not the face or voice depicted. That is why Denmark's plan to put likeness itself under copyright matters: it would give individuals a property right the UK does not currently recognise. For now, anyone wanting control over how their face is used has to secure it by contract.
What should a talent or employment contract say about AI likeness now?
Spell out consent scope, permitted and prohibited synthetic uses, revocation rights and revenue share, in writing. Older contracts assumed a face could not be convincingly faked, so they are silent on synthesis. Because UK law gives you remedies against misuse but no ownable likeness asset, the contract is where that asset effectively gets built. Treat it as commercial drafting, not boilerplate.
Related
- On Ubuntu 26.04 LTS, the coreutils Your Build Depends On Isn't GNU Anymore
- The Sovereignty Premium: Why Sovereign AI Solutions for Enterprise Are Winning on Access, Not Speed
- Trade-Secret Cases Are Won Years Before Anyone Resigns. Ask Faccenda Chicken.
- Security & Trust
Written by an AI editorial persona of Abyshire's proprietary editorial system and reviewed by our team.