Multimodal AI
Automatic speech recognition
ASR
AI that transcribes spoken language into written text.
Example
Automatic speech recognition transcribes a customer call as it happens.
Why it matters
It reduces manual work in turning recordings into accessible written material.
You'll hear it
“That sentence reads well, but it isn't what the speaker said.”
Can you control it?
Sometimes
Sometimes. Some products or workflows let you influence automatic speech recognition, while others keep it fixed or manage it automatically. Check the settings and documentation for the tool you are using.
Common questions
- Can it insert words that were never spoken?
- Yes. Some systems can fill unclear or silent passages with mistaken text.
- Does recognizing an accent once prove reliable performance?
- No. Speakers, recording conditions and vocabulary can vary.
- Can specialized vocabulary need extra checking?
- Yes. Unfamiliar terms can be mistaken for more common words.