September 10, 2025Open Access

Prosodic cues strengthen human-AI voice boundaries: Listeners do not easily perceive human speakers and AI clones as the same person

View Full Paper

Ask AI

Bookmark

Listeners consistently rated AI speech as less humanlike, regardless of prosody, highlighting prosodic expressiveness importance.
In identity discrimination tasks, listeners showed ~54% accuracy with matching prosody and ~36% with inconsistent prosody.
Bayesian modeling indicated nonlinear effects of acoustic distance on listener performance, supporting direct matching mechanisms.
Concerns about premature voice clone detection may need reevaluation, given the significant within-speaker identity variations.

Discussion

No takes yet. Share an insight, caveat, or question.

Experiment reveals AI clones are perceived less humanlike, suggesting prosody aids identity discrimination.

Chen et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark