Listeners perceive cloned voices as different from originals 41% of the time
When presented with cloned voices paired with real recordings from the same speaker, listeners disagreed with the speaker label 41% of the time, perceiving them as different people.
Models trained using speaker labels aligned better with listener perception than self-supervised models, though a model with a 12% error rate on speaker labels matched listener alignment as well as models with error rates below 2%.
Sources behind this report
- arXiv
Research · abstract · CC0 1.0
Research manuscript in arXiv. Peer-review status is not recorded here; consult the linked version and publication.
Source published
Read by the newsroom
Report history
No earlier revision dates are retained here. This does not establish that the report has never changed; read any dated updates and corrections.
This brief was made by AI and checked against its cited public sources. How the newsroom works