Voice and likeness consent for AI

Separate from copyright, using an identifiable person’s voice or likeness to train or generate can require that person’s consent under publicity and emerging likeness laws.

Why it matters to a buyer

A copyright licence to a recording does not by itself grant the right to clone the speaker’s voice. This is the most regulation-heavy gap in the market.

Why it matters to a data owner

Owners of recordings featuring identifiable people need consent covering AI use specifically before licensing.

Current legal status

Publicity rights are state-law in the US and vary widely; new statutes (Tennessee’s ELVIS Act, 2024) and the proposed federal NO FAKES Act specifically address AI voice/likeness. [Re-check current status.]

What fiund does about it

Where people are identifiable, fiund handles voice and likeness consent separately from the copyright licence.

← All rights & provenance guides