Six defendants. One legal theory the AI industry hasn’t fully mapped yet.
Voice actors filed a proposed class action alleging that Amazon, Apple, Google, Meta, Microsoft, and Nvidia harvested vocal performances to train AI voice models without authorization, according to MediaPost’s reporting. The complaint invokes the Illinois Right of Publicity Act and federal copyright law. Plaintiffs allege unauthorized commercial use of their voice likenesses, a theory that operates differently from the text-copyright claims currently working through federal courts.
The legal theory matters. The Right of Publicity protects a person’s name, likeness, and voice from commercial use without consent. It’s a state-law claim, and Illinois has one of the country’s stronger Right of Publicity statutes. The AI training data cases that have moved furthest in courts, the input-copying theory tested in Nazemian v. NVIDIA, the CMI removal claims in publishers v. Meta, rely primarily on federal copyright frameworks. This suit adds a parallel track. The Right of Publicity doesn’t require plaintiffs to prove the defendant copied a protected work. It requires showing that the defendant commercially used the plaintiff’s voice likeness without consent.
Voice Actor Class Action, Party Positions
Plaintiffs allege the defendant companies ingested voice recordings to train AI systems, including voice assistants, audiobook narration tools, and text-to-speech products, without permission. The complaint claims this constitutes unauthorized commercial exploitation of the plaintiffs’ professional performances. What the defendants actually used, and what their training pipelines contained, isn’t established by this complaint. These are allegations.
The defendant list is notable for its breadth. Amazon (Alexa), Apple (Siri), Google (Assistant and related products), Meta (voice AI in AR/VR and social platforms), Microsoft (Cortana, Azure speech services), and Nvidia (Riva speech AI platform) all have commercial voice AI products. That all six appear in a single complaint suggests plaintiffs are framing this as an industry-wide practice, not a single-company failing.
The catch is that state Right of Publicity claims and federal copyright claims don’t always coexist cleanly in litigation. Federal copyright law can preempt state claims that are “equivalent” to copyright, defendants will almost certainly argue preemption. Whether Illinois’s Right of Publicity survives that challenge is an open legal question that will shape this case’s viability before it ever reaches the merits.
Unanswered Questions
- Will the court allow Illinois Right of Publicity claims to survive federal copyright preemption arguments?
- Do existing AI voice model training practices constitute 'commercial use' of voice likeness under Illinois statute?
- Does this theory extend to companies using publicly available recordings (podcasts, audiobooks) vs. contracted performance recordings?
What to watch
The preemption briefing. If defendants move to dismiss on federal preemption grounds, the court’s ruling will signal whether Right of Publicity claims can survive as a standalone theory in AI training cases, or whether the only viable path runs through federal copyright.
TJS synthesis: Every AI copyright case that survives a motion to dismiss teaches the industry something about where the legal exposure actually sits. The Right of Publicity theory, if it clears preemption, means voice AI developers face potential liability under 50 different state statutes, not just federal copyright. The real question isn’t whether this particular case succeeds. It’s whether the preemption argument fails, opening a second legal front that the industry hasn’t priced in.