Arnold Peter
Founding Partner at Peter Law Group
Date: September 1, 2026
Originally published on LinkedIn
Summary: Arnold Peter examines how AI voice cloning and digital likenesses are reshaping the relationship between technology and personal identity. The article explores the opportunities for performers to license digital replicas, the risks facing workers with less bargaining power, and the importance of meaningful consent, compensation and contractual limits. Drawing on Hollywood’s voice-cloning debate and allegations in the firm’s Avatar litigation, Peter argues that established legal principles can help preserve human control as new technologies replicate voices and faces.
I have written quite a bit recently about artificial intelligence, particularly where technology collides with law, creativity and human judgment. The more I study this subject, the more convinced I become of something that initially sounds contradictory: AI is both far more transformative-and far less unstoppable-than many people believe.
The technology is moving extraordinarily fast, but technology does not operate in a vacuum. Courts, legislatures, unions, contracts and ultimately human beings are beginning to establish boundaries around what AI can and cannot do. Nowhere is that tension becoming more visible than in Hollywood.
A fascinating new Los Angeles Times article examines the rapidly developing battle over AI voice cloning, and it raises a question much larger than whether an actor should license his or her voice: what happens when technology can separate a person’s identity from the person?
The Voice Actor May No Longer Need to Be in the Room
For generations, an actor’s voice was inseparable from the actor. If you wanted James Earl Jones’ voice, you needed James Earl Jones; if you wanted an actor to narrate an audiobook, commercial or documentary, that actor had to perform the work.
AI changes that equation. As the Times reports, established performers are beginning to license digital replicas of their voices: Michael Caine has licensed his voice through ElevenLabs, while Matthew McConaughey has used the technology to produce a Spanish-language version of his newsletter. Meanwhile, voice marketplaces are emerging in which licensed replicas can generate income for their owners. For a successful performer, that can look like an extraordinary business opportunity: record once, license repeatedly, and earn money while sleeping.
But move a few rungs down the economic ladder and the same technology looks very different. Freelance voice actors are watching the smaller jobs that traditionally sustained careers-commercials, audiobooks, online courses, YouTube videos and corporate narration-begin to disappear, and some performers are being offered money to provide the recordings necessary to train AI systems that could eventually compete with them. That creates a remarkable economic proposition: we will pay you today to help us build something that may eliminate your job tomorrow.
This Is Not Really a Fight Between Actors and AI
It is tempting to frame this debate as humans versus machines, but I think that misses the point. AI is a tool, and the important questions are about the rules governing that tool. Who owns a voice? What exactly does an actor license when consenting to the creation of a digital replica, for how long, and for what purposes? Can the voice be used in a political advertisement, pornography, a product endorsement, or a foreign-language production the actor has never seen? Can the license be revoked, and does the actor approve individual uses? What happens when the actor dies? And perhaps most importantly: can an AI company train a model on someone’s voice without that person’s meaningful consent?
Those are not primarily technological questions. They are questions of contract, intellectual property, right of publicity, labor law, privacy and human autonomy-and that is where lawyers, courts and legislatures enter the picture.
The Law Is Beginning to Catch Up
This fits a pattern I have been writing about for some time. The technology side of AI sometimes appears almost boundless, but on the legal side, we are increasingly seeing institutions impose human accountability on systems that otherwise have none.
SAG-AFTRA has already negotiated significant protections involving digital replicas. According to the Times, current protections include notice and compensation requirements when studios create and use replicas, payment for certain subsequent uses and restrictions on deploying replicas during strikes. But there is an enormous hole in that protection. The National Association of Voice Actors estimates that approximately 80% of voice-acting work is freelance and therefore outside union protections, and that may be where the real legal battle develops.
A famous actor represented by sophisticated agents, lawyers and a powerful union can negotiate the economic value and permissible uses of a digital replica, while a 24-year-old actor desperate for a first paying job has considerably less bargaining power. The technology may be identical, but the ability to negotiate its use is not.
Consent Is Going to Become One of the Most Important Words in AI Law
This is why I believe consent will become one of the central concepts in the emerging law of artificial intelligence-not consent buried on page 47 of a click-through agreement, not perpetual consent hidden inside boilerplate, and not “consent” obtained from someone who does not reasonably understand that six hours of recordings could create a digital performer capable of functioning indefinitely without them.
What we need is something closer to informed consent: What is being created, and how will it be used? Who controls it, and how long can it be used? Can it train other models or be sublicensed? What compensation will be paid, and can consent later be withdrawn?
Consent, however, does not operate in isolation, and it should not be left entirely to private bargaining between unequal parties. A union representing performers may negotiate collective terms that prescribe or restrict what its members can be asked to consent to—capping the duration of a replica license, barring certain categories of use, or guaranteeing compensation for downstream exploitation. And a legislature may impose baseline requirements that cannot be bargained away at all: minimum protections that survive even where a young performer, desperate for a first paying job, would otherwise sign them away in perpetuity. In other words, some forms of consent should simply be unavailable, no matter how willingly they are offered, because the point of these rules is to protect the person with the least leverage, not to validate whatever the most powerful party can extract.
The Times article illustrates why this matters. One performer who licensed an AI version of her voice described subsequently hearing that voice in projects she had not anticipated-and that should concern anyone, not just actors.
From Voice Cloning to Facial Likeness: The Avatar Litigation
Today’s dispute involves voices, but tomorrow it involves faces, movements, writing styles, musical performances and eventually increasingly sophisticated replicas of entire human identities. I have seen this dynamic firsthand.
My firm recently filed a lawsuit in the United States District Court for the Central District of California on behalf of our client, actress Q’orianka Kilcher, alleging that filmmaker James Cameron used her facial features—her lips, chin, jawline and overall facial structure—without her authorization as the foundation for the appearance of Neytiri, the Na’vi lead of the multibillion-dollar Avatar franchise. According to the complaint, Mr. Cameron drew on a photograph of Kilcher taken when she was fourteen, and her likeness was then captured in production art, sculpted, scanned and integrated into the digital pipeline that generated the character—all without her knowledge, consent, credit or compensation.
What makes our case instructive is precisely that the technology is new but the legal questions are not. The dispute arises in a technologically driven format—a biometric likeness extracted and industrialized through a modern production pipeline—yet it turns on principles the law has long recognized: informed consent, the right of publicity, and fair compensation for the commercial exploitation of a person’s identity. In that sense, the Avatar litigation is less a departure from established law than an application of it to a new set of facts, and it previews the questions that voice cloning, and every other form of digital replication, will force courts to confront.
We Should Not Try to Stop AI
I am not in the camp that believes the solution is prohibiting AI. That battle is already lost-and I am not sure we would want to win it, in any regard.
Used responsibly, AI can democratize creativity. A small filmmaker may suddenly afford sophisticated dubbing. An independent producer can accomplish things previously available only to major studios. Content can be translated and localized almost instantaneously. Performers themselves may be able to monetize their identities in entirely new ways.
Those are real benefits. But innovation cannot mean that whoever develops the most powerful technology automatically acquires the weakest person’s rights. The objective should not be to stop AI. It should be to establish rules that allow AI and human creativity to coexist. That means meaningful consent, transparency, compensation, contractual limitations, protection against unauthorized replication and, above all, continued human accountability.
The Bigger Question
There is something almost philosophical underneath this dispute. For centuries, technology primarily replaced human labor. Machines replaced physical strength. Computers replaced calculation. Software automated clerical work.
Generative AI is different because it increasingly replicates things we associate with human identity: a person’s voice, a person’s face, a person’s writing, a person’s artistic style and eventually, perhaps, a remarkably convincing simulation of the person themselves. That is why the Hollywood voice-cloning dispute deserves attention well beyond Hollywood. We are beginning to determine where the boundary lies between technological innovation and human identity.
I suspect the most important question of the AI era will not be what artificial intelligence can do; technology will answer that question for us. The much harder question-and the one that lawyers, courts, creators, businesses and society will have to answer-is what we should allow it to do, and under what conditions.
This is precisely why the conversation has moved well beyond courts and contracts. In May 2026, Pope Leo XIV devoted his first encyclical, Magnifica Humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence, to exactly this question, framing the protection of human dignity as the defining challenge of the AI age. An encyclical is a formal, authoritative statement from the Pope-typically addressed to the world’s bishops and the faithful-setting out the Church’s teaching on a matter of major moral or social importance.
His central insight resonates far beyond any single faith tradition: technology is not inherently evil, but neither is it neutral, because it takes on the characteristics of those who devise, finance, regulate and use it. He signed the document on the 135th anniversary of Pope Leo XIII’s Rerum Novarum, the encyclical that responded to the upheavals of the Industrial Revolution and shaped modern thinking on labor and technology. The parallel is deliberate and telling: just as the last great transformation demanded new rules to protect the worker, this one demands rules to protect the person. When an institution two thousand years old feels compelled to weigh in, it is a signal that the boundary between technological innovation and human identity is not a niche legal problem but a question that every part of society-sacred and secular alike-will have to help answer.
The Los Angeles Times article is worth reading because the voice actors fighting over this issue may simply be among the first workers confronting a question that eventually reaches nearly every profession.
Frequently Asked Questions About AI Voice Cloning and Identity Rights
What is the article’s central argument?
Peter argues that the key question is not simply what AI can replicate, but under what conditions it should be allowed to do so. He emphasizes meaningful consent, compensation, contractual limits and human accountability.
What questions does the article raise about digital-replica licenses?
The article asks what is being created, who controls it, how long a license lasts, which uses are permitted, and whether the replica can be sublicensed or used to train other models. It also raises compensation, approval of individual uses and withdrawal of consent.
Why does bargaining power matter in the article?
Peter contrasts established performers with agents, lawyers and union support with freelance actors seeking their first opportunities. He argues that the ability to negotiate meaningful limits can differ sharply even when the technology is the same.
How does the article connect voice cloning with facial likeness?
Peter discusses the firm’s lawsuit on behalf of Q’orianka Kilcher as an example of identity-rights questions arising in a digital production process. The article describes allegations concerning the use of her likeness; it does not present those allegations as a court finding.
Does the article advocate banning AI?
No. Peter describes potential benefits such as affordable dubbing, translation, localization and new licensing opportunities. He argues for rules that allow those benefits while protecting people against unauthorized replication and exploitation.
Why does this discussion extend beyond entertainment?
The article argues that AI increasingly replicates aspects of personal identity, including voices, faces and creative expression. Peter therefore sees consent and control over digital identity as questions that may affect workers and businesses well beyond Hollywood.
