Right to likeness and voice
How it works
Protection works by granting a person control and consent rights: disseminating their likeness or voice without permission is unlawful and gives rise to claims. Enforcement mechanisms include civil actions (injunctions, takedowns, damages, compensation for harm), court orders, and — in newer statutes — liability extending to AI-generated digital replicas and to distributing tools that produce them. Sector rules such as the EU AI Act impose transparency duties, requiring generated or manipulated (deepfake) content to be labelled. Scope varies by jurisdiction: from civil-law personal goods, through state-level rights of publicity in the US, to proposed federal legislation.
Problem solved
Addresses the unauthorized use of a person’s identity — their face, body, voice and style — especially now that generative AI can produce convincing digital replicas without consent. It guards against commercial exploitation, disinformation, fraud and harm to dignity (e.g. deepfake pornography, fake endorsements, voice clones).
Components
Using a person’s likeness or voice generally requires their consent; consent can be scoped and licensed (e.g. for advertising).
US doctrine treating the commercial value of identity as a property right, often descendible; the term was coined by Judge Jerome Frank in 1953.
An AI-generated, realistic reproduction of a person’s voice or likeness; addressed directly by the ELVIS Act and the proposed NO FAKES Act.
Generative techniques that synthesize faces and voices, making likeness and voice protection practically urgent.
Diverse sources: personal goods (Polish Civil Code arts. 23/24, Copyright Act art. 81), US state laws, the EU AI Act (deepfake transparency), and proposed federal legislation.
Evolution
Judge Jerome Frank introduces the concept in Haelan Laboratories v. Topps Chewing Gum.
Appeals court recognizes that a distinctive voice is protectable against imitation in advertising (849 F.2d 460).
A track with cloned Drake and The Weeknd vocals (Ghostwriter977) highlights artist voice cloning.
Ensuring Likeness, Voice, and Image Security Act signed by Gov. Bill Lee on 21 March 2024 — first state law protecting voice against AI.
OpenAI pulls the “Sky” voice after claims it resembled the actress’s voice (May 2024).
Proposed federal digital-replica bill (draft 2023, introduced 2024, reintroduced 2025) establishing a federal right of publicity.
Obligations to label generated and manipulated (deepfake) content apply from 2 August 2026.