01
Factual background & dispute
- The voice actors accepted limited-scope recording assignments through a freelance platform.
- The plaintiffs subsequently discovered that synthetic voices closely mimicking their own were marketed in AI voiceover products.
- The complaint asserted claims under copyright, trademark, contract, right of publicity, and consumer protection laws.
02
Core issues & judicial focus
- Which distinct legal protections apply to a voice itself, the sound recordings, and synthetic voice clones
- Whether a limited-use voice assignment extends to AI model training and commercial voice synthesis products
- How training-stage copying and synthetic model outputs should be evaluated separately
03
Judicial finding & holding
- Sage’s copyright claim based on use of original recordings survived, alongside contract, New York Civil Rights Law and consumer-protection claims. This was not copyright in voice characteristics themselves.
- Training-related copyright claims were dismissed with leave to amend for insufficient facts; specified trademark and other copyright claims were dismissed. The court did not find training lawful or fair use on the merits.
- Lovo’s 27 May 2026 notice reported Chapter 7 case 26-11249. The defendant’s notice was not a stay order, and subsequent procedure remains unverified.
Dates & procedural record
- ECF45 ruled on the motion to dismiss.
- The defendant’s ECF65 notice reported a Chapter 7 filing.
- The reported stay order and subsequent procedural status remain unverified.Online sources; no specific document reviewed.
04
Practical risk implications
01Voice licensing agreements should delineate distinct provisions for master recordings, vocal characteristics, AI training rights, and synthetic outputs.
02AI voice vendors must maintain provenance records and licensing documentation for source recordings and derivative voice models.
03Downstream procurers should secure clear sublicensing scope, representations on personality rights, and commitments for ongoing takedown support.