UK Actors Demand Legal Ownership of Their Voices as AI Cloning Scams Spread

Around 80 British actors, audiobook narrators, singers and comedians have signed an open letter to Prime Minister Andy Burnham demanding that the law recognise a person’s voice as a protected part of their identity — something that, unlike a name, a face in a photograph or a registered trademark, has no dedicated standing in UK statute.

The campaign, Save Our Voices Now, went public on 28 August 2026 with backing from Nicola Coughlan, Hugh Bonneville, Matt Lucas, Luke Evans, Siobhán McSweeney, Pearl Mackie, Niamh Cusack, the comedian Jen Brister and the singer Sandi Thom. Alongside the letter, organisers launched a petition on the gov.uk site asking that ownership of one’s own voice be written into the statutory rights of every person in Britain — not only of working performers.

“My voice is unique. It’s mine. It’s part of my identity. So is yours. That’s worth speaking up about,” said Bonneville, who added, drily, that the quote was handwritten and not created by AI.

The campaign’s central statistic is not about lost residuals but about fraud: Save Our Voices Now says 28 per cent of UK adults report having been targeted by a voice-cloning scam. The figure is the campaign’s own, but it anchors an argument that a performers’ grievance is really a consumer-protection problem wearing a celebrity face.

The campaign and the gap it points to

Save Our Voices Now was founded by the actor Peter Caulfield and co-founded by the audiobook narrator Alice Sockett, who has described the issue as an existential threat to her industry and says reports of unauthorised copies are mounting weekly.

“In just three seconds, with the right AI system, your voice can be cloned, stolen and reinterpreted without your knowledge,” Caulfield said. “This isn’t fiction. It’s happening right now. The bottom line is the government is allowing AI companies to steal and profit from artists’ voices without their consent. They are also allowing criminals to use AI to clone people’s identities and threaten everyone. We need to use our real voices now to stop it.”

McSweeney, who starred with Coughlan in Derry Girls, put the case in terms of pace rather than principle. “I’m deeply worried that legislation, contracts, working practices and ethics, and artists’ rights are not being considered or protected as quickly as AI technology is progressing,” she said. “This is an existential problem and needs to be controlled legislatively as soon as possible.”

There is no single British statute covering a synthetic likeness. A claimant today has to assemble a case out of passing off, defamation, data protection law and, occasionally, copyright in one specific recording — none of which were drafted with machine learning in mind. The performers’ union Equity has filed claims on behalf of more than 20 members whose voices were allegedly used without permission, and now supplies members with a template contract and a template takedown notice developed with the intellectual property academic Dr Mathilde Pavis.

The government has so far offered process rather than legislation. A spokesperson said ministers “value and protect human creators, and that is why we have announced we will launch a consultation on how we address these harms while protecting legitimate innovation,” with officials saying they will examine whether it would be beneficial to introduce a new digital replica or personality right. Signatories read a consultation as another delay — the second in a year, after ministers postponed planned AI copyright changes in March following a hostile response from the creative sector.

Analysis

The demand here is not a copyright claim, and that distinction is the whole argument. Copyright attaches to a fixed work — a particular recording, a particular performance. A cloned voice usually infringes nothing so tidy: the output is new audio that no one else authored, generated by a system trained on material that may have been lawfully published. What the signatories want instead is a personality right, treating identity itself as the protected thing. France, Italy, the Netherlands and Denmark have long recognised something along those lines. Britain never has.

Denmark has gone furthest. Culture Minister Jakob Engel-Schmidt announced reforms in June 2025 that would amend Danish copyright law to give citizens legal ownership of their own face, body and voice, with the right to demand removal of unauthorised AI content, to seek compensation, and with protection extending up to 50 years after death. Parody and satire are carved out. Britain has no equivalent bill and no deadline.

The United States has moved through the states rather than Washington — New York’s digital replica contracts law and Tennessee’s ELVIS Act are the clearest examples — producing protection that varies by postcode and largely runs through contract law. The EU’s AI Act handles deepfakes mainly through transparency duties, not ownership. Three jurisdictions are converging on one problem from three incompatible directions, while the underlying material sits on servers that respect none of those borders.

Publicly available audio of a professional narrator or broadcaster is effectively infinite, which is why voice artists rather than screen stars are pressing hardest. But the fraud statistic points elsewhere: at ordinary people with no agent, no union template and no realistic route to court. There is also a counterweight the campaign acknowledges. Cloned voices restore speech to people who have lost theirs to conditions such as motor neurone disease — a use signatories do not want banned. Any statutory right must therefore separate consent from mere availability, which is precisely the drafting problem a consultation is meant to resolve.

Watch three things: whether the promised consultation is given a firm start date and terms of reference this autumn; whether the gov.uk petition clears the 10,000-signature threshold that compels a written government response, or the 100,000 that triggers a Westminster Hall debate; and whether Denmark’s bill completes passage and becomes a template other EU states copy. The nearer commercial risk is procurement: synthetic voice is already routine in contact centres, e-learning and advertising, and liability may not stop with the vendor that trained the model.