This study investigated the intelligibility and processing characteristics of Mandarin sentences synthesized based on a large-scale language model (LLM), compared to natural speech, specifically for deaf adults with cochlear implants (CIs).
No actionable change for current clinical practice; findings are exploratory and relevant primarily to researchers developing AI-based speech tools for Mandarin cochlear implant users.
As AI-generated speech becomes increasingly common in daily life, understanding how cochlear implant users perceive it is critical for designing accessible AI interfaces and future rehabilitation tools.
- 01Tests intelligibility of large language model (LLM)-synthesized Mandarin speech for cochlear implant users.
- 02Deaf adults with CIs were the study population.
- 03Addresses a gap in how emerging AI speech synthesis interacts with CI hearing.
- 04Published in Journal of Speech, Language, and Hearing Research.
- 05May inform future AI-based auditory rehabilitation or assistive technology design.
LLM-synthesized Mandarin speech can be perceived and processed by deaf adults with cochlear implants.
studypartially supported- PMID
- 42372716
- DOI
- 10.1044/2026_JSLHR-25-00295.
- Journal
- Journal of Speech, Language, and Hearing Research
- Publication type
- research_article
- Evidence level
- 4
- Population
- Deaf adults with cochlear implants, Mandarin-speaking
- Intervention
- Large language model (LLM)-synthesized Mandarin speech perception
Primary outcomes
Intelligibility of LLM-synthesized Mandarin sentences; Speech processing performance among CI users