The photograph is flawless. The endorsement looks genuine. But the individual at the centre of it was never approached, never photographed, and never gave consent. Welcome to the age of generative AI, where a synthetic image indistinguishable from reality can be conjured in seconds and shared with millions before anyone thinks to ask whether the person depicted ever gave their consent.
For athletes, entertainers, public figures, and other high profile individuals, this is no longer a hypothetical risk. It is a live, recurring problem, and one that exposes a significant gap in how English law protects a person’s image. Deepfake videos, AI-voice replication, and synthetic images are already circulating at scale across social media. Recent reports that AI-generated fake Facebook news pages have used the image and likeness of actress Lisa Riley without her knowledge or consent illustrate the point sharply. Meanwhile, more than 80 performers, including actors Matt Lucas, Hugh Bonneville and Nicola Coughlan, have written to the Prime Minister demanding legislation that would give every person in the UK a legal right to own their voice. Audiobook narrator Alice Sockett, who co-founded the campaign, has called voice cloning an existential threat to the entire industry and has objected to her voice being in the public domain completely unprotected.
Image right? There’s no such thing
The uncomfortable truth is that there is no standalone “image right” in English law. Individuals whose likeness is used without their consent must instead piece together a claim from a constellation of legal doctrines that were never designed with generative AI in mind. As Dr Mathilde Pavis has observed, “The law comes close, but it doesn’t fully protect your voice, your face or your identity.” That may be about to change: a government Report on Copyright and Artificial Intelligence published in March 2026 acknowledged that “existing protections do not give most individuals meaningful control of their image or voice or access to realistic remedies” and stated that the government proposes “to explore options that address these risks, while promoting growth an innovation. This will include considering whether a new personality right may be appropriate”.
A consultation around the issue of digital replicas (including the unauthorised use, creation and distribution of AI-generated content that reproduces an individual’s image, voice or likeness) was due to be launched this summer but has not yet opened. So, for now, individuals must work with the tools the law currently gives them.
What are image rights?
Image rights is the shorthand for a cluster of rights that can be leveraged by individuals to control the exploitation of their name, likeness, voice or other personal attributes. The underlying principle is that an individual should be able to prevent unauthorised use of their name, face, voice and other physical characteristics, and that others should not be free to capitalise on these very personal attributes without consent.
Without a standalone image right in English law, individuals must rely on a combination of intellectual property law, contract law, data protection, and common law, none of which were developed with generative AI in mind. Taken together, these protections offer limited practical redress for those whose likeness is used without consent.
This is the first in a series of articles exploring how the law is responding to rapid development in AI technologies versus the need to protect an individual’s right to control use of their image. Our next article will turn to lessons from other jurisdictions that have already grappled with the question of image rights and what their approaches might offer as a model for reform closer to home. A further article will turn to some practical scenarios, looking at how the UK framework may currently operate when put to the test.
The current framework
Previously we’ve examined how the current legal framework in England and Wales operates regarding AI-generated content such as deepfakes, focusing on the protections available through Intellectual Property and Data Protection laws.
IP
In summary, from an IP perspective, the two principal options to challenge unauthorised AI-generated content are passing off and copyright infringement. Both of these face significant limitations. A passing off claim requires the individual to demonstrate (i) goodwill through commercial exploitation of their image, (ii) a misrepresentation capable of deceiving the public, and (iii) resulting damage. While well-known individuals may be able to rely on passing off in certain scenarios, this avenue is less accessible for lesser-known artists and creatives, or the general public. More fundamentally, passing off protects goodwill, not dignity or autonomy. It offers little for the individual who finds their face grafted onto degrading, misleading or defamatory content where there is no commercial endorsement to misrepresent.
Copyright, meanwhile, protects the work (i.e. the photograph, film or recording) not the person who appears in it. Copyright ownership will usually vest in the photographer or filmmaker (subject to any assignment or contractual arrangement to the contrary). The individual whose likeness has been reproduced will rarely be the rights holder and so will lack standing to bring a claim in their own right. Even where a copyright owner is willing to act, establishing infringement presents its own difficulties: the composition of AI-generated content may draw on many different sources, making it hard to show that a substantial part of any single protected work was used in the AI output, or in training the model.
The March 2026 report on Copyright and AI confirmed that the government will not be introducing changes to copyright law until it is confident that such measures will meet its objectives for the economy and UK citizens. The report indicated that any reform “must ensure that right holders can be fairly rewarded for the economic value their work creates, and that they are protected against unlawful and unfair use of their work”. Reform must “also ensure that AI developers can access high-quality content”. The report concluded that there is currently no consensus on how these competing objectives are to be achieved.
In relation to digital replicas, the report acknowledged that this is a growing area of concern for creative industries, and that the existing legal framework may not be sufficiently robust to deal with the issue of unauthorised digital replicas. Stakeholders supported enhanced protections for a person’s image and voice, but no single view of what form this protection should take emerged from the consultation.
Data protection
Data protection law approaches the problem from a different angle. Because generative AI systems must ingest and process images or recordings of identifiable individuals in order to produce synthetic content, that activity amounts to the processing of personal data. Where facial features are analysed for the purpose of identifying a person, for example as part of the creation of AI-generated content that replicates their likeness, the stricter rules for special category biometric data may also apply (including separate conditions for processing). Any such processing requires a lawful basis. If the basis relied upon is legitimate interest, the processing must survive a balancing exercise that weighs the creator’s interest against the individual’s rights, including the potential for reputational harm, distress, or loss of control over their own image. While data protection law may provide individuals with regulatory and legal remedies after the event, including the ability to complain to the ICO or seek compensation, it is not generally designed to prevent AI-generated content from being created or disseminated in the first place.
The Riley case also highlights the question of platform responsibility. Where AI-generated content is disseminated through social media at speed and at scale, the practical burden of identifying, reporting and requesting removal often falls on the individual whose image has been misappropriated, rather than on the platform hosting it.
While copyright, passing off, trademarks, contractual rights, regulatory enforcement and data protection may each offer a degree of protection in certain situations, none, whether alone or in combination, delivers the clarity or comprehensiveness that a standalone image right would afford.
What comes next?
The practical question now shifts to what a workable framework could contain. The UK would not be the first jurisdiction to act. In the United States, the pending NO FAKES Act proposes a federal right over voice and visual likeness, while Germany has recognised personality rights as a feature of its civil law since the 1950s. Denmark has announced proposed amendments to its copyright legislation to provide individuals with more protection over the use of their identities.
Balancing innovation and the legitimate commercial use of AI against an individual’s interest in controlling their own image will require careful calibration. The anticipated consultation on digital replicas will be a critical test of whether the existing framework of piecemeal protections could be replaced with something more targeted and coherent.
Our content explained
Every piece of content we create is correct on the date it’s published but please don’t rely on it as legal advice. If you’d like to speak to us about your own legal requirements, please contact one of our expert lawyers.