Fact · Transcription
A 2025 study of non-native English speech found that speech recognition systems vary significantly in how they handle filler words, repetitions and revisions.
Checked against its source on October 17, 2026
In the source's own words
“finding significant variation in performance across systems and disfluency types”
Automatic Speech Recognition for Non-Native English: Accuracy and Disfluency Handling, Michael McGuire, arXiv, 2025, Abstract.
Open the original- What it measures
- The study compared several recognizers on read speech and on spontaneous narrative speech, and scored how each handled disfluencies.
- What it does not show
- The abstract reports the finding without per-system numbers here. It covers non-native English speakers and a set of commercial and open systems that changes quickly.