Fact · Languages and accents
A review chapter on pronunciation training says that speech recognition systems, such as those on Siri or Alexa, treat a learner's mispronunciations as noise, because their goal is to transcribe the sounds the speaker intended.
Checked against its source on October 17, 2026
In the source's own words
“An ASR system (such as those on Siri or Alexa) treats mispronunciations as noise; that is, the goal of ASR is to transcribe the intended sequence of sounds.”
Speech Synthesis and Pronunciation Teaching, Waris Quamer et al., The Encyclopedia of Applied Linguistics, 2nd edition (in press), author preprint, 2024, Section 2, Explicit feedback via speech recognition.
Open the original- What it measures
- The chapter contrasts recognition with mispronunciation detection, which is built to find the errors. It describes the design goal of the systems, and it notes that the recognizers are trained on speech from many speakers regardless of accent.
- What it does not show
- This describes how general recognizers are designed, not how a particular app responds to your speech in practice. The chapter does not measure any app, so test the app yourself with a sentence you know you mispronounce.