Seeing an iconic artist appear to sing live, reconstructed through voice synthesis technology, is genuinely impressive. The emotional response is real. The sense of witnessing something that until very recently seemed impossible carries its own weight.
The technology has crossed a threshold. Platforms like ElevenLabs can now create voice replicas of sufficient quality that listeners who are not specifically listening for artificiality may not detect it. Artists like Matthew McConaughey and Michael Caine have already authorized commercial use of their voice through formal licensing arrangements. And in several cases, deceased artists have appeared to sing in contexts where no such authorization existed or was sought.
The question the industry has not yet answered clearly is: when a voice is recreated through AI, what legal and commercial framework governs it?
The Rights That Come Into Play
At least three distinct categories of rights are relevant when a voice is recreated using AI technology.
The first is intellectual property in the original recordings. If a voice synthesis model was trained on existing recordings of an artist, the copyright in those recordings belongs to whoever holds it, typically a record label or the artist's estate. Using those recordings to train a model without a license is the same category of infringement as any other unauthorized use of protected audio.
The second is what the legal system calls NIL rights: Name, Image, and Likeness. These rights protect the commercial value of an individual's identity. They are well established in sports and advertising law, and they are increasingly being applied to the creative industry as AI makes unauthorized reproduction of identity commercially viable at scale.
The third, and the one most directly relevant to Collective Management Organizations, is the right of public communication. When a recreated voice is performed or broadcast in a commercial context, including on streaming platforms, at live events, or in audiovisual productions, the relevant rights for public performance or communication typically fall within the mandate of collecting societies.
Who Authorizes and Who Gets Paid
In the case of a living artist, the authorization question is relatively straightforward. The artist can choose to license their voice, as McConaughey and Caine did through ElevenLabs. That license governs the terms of use, including what commercial applications are permitted, what compensation is due, and what quality controls apply to the output.
When the artist is deceased, the situation is more complex. Rights in recordings are typically held by a label or an estate. NIL rights may or may not pass to heirs, depending on the jurisdiction. And the question of whether a collecting society is receiving and distributing public communication royalties for uses of a recreated voice often goes unanswered simply because there is no established process for handling it.
The Gap Between What Technology Can Do and What Legal Frameworks Cover
The honest assessment is that the legal frameworks governing AI voice recreation are incomplete in most territories. Courts and legislators are working to close the gap, but the technology moves faster than the law.
This is not a reason for Collective Management Organizations to wait. CMOs have historically been the institutions that build the practical infrastructure for rights management in areas where the law is clear in principle but complex in execution. The principle here is clear: creators and their estates should be compensated when their identity and voice are used commercially. The execution requires building monitoring systems, establishing licensing frameworks, and developing the processes for distribution.
Several jurisdictions are beginning to address this legislatively. The state of Tennessee in the United States passed the ELVIS Act specifically to protect recording artists from unauthorized AI vocal replicas. The European Union's AI Act includes provisions relevant to synthetic media and identity rights. These are early steps toward a framework that will eventually need to cover every territory where CMOs operate.
The Role of Collecting Societies Going Forward
Organizations that manage rights for performing artists are positioned to extend their mandate into this space more naturally than any other type of institution. They already represent the rights holders. They already have the collection and distribution infrastructure. They already have relationships with platforms and broadcasters.
Building the capacity to monitor AI-generated voice content, to assert rights on behalf of members whose voices are recreated without authorization, and to negotiate licensing frameworks with AI companies for authorized uses is a logical extension of what collecting societies already do.
The question is whether organizations will build that capacity proactively or reactively, and whether they will do it in time to shape the frameworks rather than simply enforce whoever else designed them.
The Labs at global.esur researches and develops solutions for the most pressing technology challenges facing Collective Management Organizations. Contact us to explore what AI voice rights mean for your members.