Fact · Dictation for work and study
Apple researchers report that end-to-end speech recognition still has a significant error rate for entity names that appear infrequently in its training data.
Checked against its source on October 16, 2026
In the source's own words
“these systems still have a significant error rate for entity names which appear infrequently in their training data”
Retrieval Augmented Correction of Named Entity Speech Recognition Errors, Ernest Pusateri and et al. (nine authors in total), arXiv, 2024, Abstract.
Open the original- What it measures
- The abstract describes end-to-end speech recognition systems, which map sound directly to text. Entity names are the people, places, products and invented names that appear in a book.
- What it does not show
- The paper gives no single error rate for everyday dictation, and its test sets focus on voice assistant requests. Treat it as evidence of the pattern, not of how often a particular app will misspell a name.
Source
Cited on
- Writing a Book by Dictation: Software and WorkflowGuide
- How Accurate Is Speech Recognition? The Real NumbersGuide
- Dictation for Academics, Researchers and ThesesGuide
- Can you dictate a novel?Question
- How do you dictate citations and references?Question
- Why does speech recognition get names and jargon wrong?Question