Fact · Speech recognition
The 2026 draft of Jurafsky and Martin's textbook says that automatic speech recognition has matured to the point where it is viable for many practical tasks, while automatic transcription by any speaker in any environment is still far from solved.
Checked against its source on October 17, 2026
In the source's own words
“ASR technology has matured to the point where it is now viable for many practical tasks.”
Speech and Language Processing, Chapter 16: Automatic Speech Recognition (draft), Daniel Jurafsky and James H. Martin, Stanford University, online manuscript of the third edition, 2026, Chapter 16, opening paragraphs.
Open the original- What it measures
- The sentence comes right after the authors say that transcription by any speaker in any environment is still far from solved.
- What it does not show
- A qualitative judgment by the authors. It gives no error rates and does not say which tasks count as practical.