Human Voices in the Age of AI

Artificial intelligence has transformed how media is created, enabling machines to replicate human speech with remarkable realism. While these technologies unlock new creative and commercial possibilities, they also raise a fundamental legal question: who owns a person’s voice once it has been cloned by AI?

The Rise of Voice Cloning

Early text-to-speech tools have evolved into sophisticated AI systems capable of recreating a person’s unique vocal characteristics including tone, pitch, cadence, and accent. These tools have legitimate uses, such as improving accessibility for people with speech impairments and enabling seamless dubbing in entertainment applications. Unfortunately, this same technology can be exploited to create deepfakes, enabling impersonation that facilitates fraud and inflicts reputational harm. Because a voice is closely tied to personal identity, its replication presents serious ethical and legal concerns.

Why Copyright Law Falls Short

Traditional intellectual property law offers little protection for the human voice itself. Copyright protects recorded works, scripts, and compositions rather than the distinctive sound of a person’s voice. As a result, AI systems can generate synthetic speech that sounds identical to an individual without copying a protected recording, exploiting a significant legal gap.

The Role of the Right of Publicity

Most legal protection for vocal identity currently comes from the right of publicity, which gives individuals control over the commercial use of their name, likeness, and voice. Courts have recognized that unauthorized imitation of a distinctive voice can violate these rights, particularly in commercial contexts. Recent decisions and legislation suggests a growing recognition that a person’s voice is a form of personal property deserving protection, even after death.

Industry Response and Contractual Protections

Performers and voice actors have been especially affected. To address the risk of unauthorized AI training and voice replication, many now rely on contractual tools such as synthetic voice or AI riders. These provisions prohibit the use of recorded performances for AI purposes without explicit consent and compensation. Regulatory efforts are also emerging, including proposed U.S. legislation and the European Union’s AI Act, which aim to impose transparency and accountability requirements on AI developers.

Final Thoughts

While AI developers may own the technology, the voice itself should remain the property of the individual. As the law evolves, personality rights and targeted legislation are increasingly filling the gaps left in traditional copyright law.