What is Consent?
The speaker knows how their recording will be used, and has signed to say so.
This is the first gate on whether speech data can be used commercially. A recording without a consent form cannot be used legally, no matter how good it sounds.
Children's speech is stricter. A guardian has to sign. That is a hard precondition on any children's speech category we take on.
Levels of consent
What a consent form permits is the first distinction. One form covers research use only; another covers commercial use, model training, or resale of the data. The scope written on the form sets the ceiling on everything downstream. Data collected under a narrow form cannot be widened later without going back to the speaker, and going back is usually impossible once a project has finished.
How it was captured is the second. A signed release is the standard. A recorded verbal statement can serve the same purpose, and it has an advantage: it shows the person was told the purpose in their own language, in their own words. A checkbox buried in an app is the weakest form, and it is the one most likely to fail a buyer's review.
Who gives it is the third. Adults consent for themselves; for children a guardian signs, and that is a hard precondition rather than a formality. For material recorded in a workplace or at an event, a company signature does not stand in for the speaker's own consent, because the person in the recording is the one whose permission is needed.
What to specify when buying
Ask to see the consent form itself, not a description of it. The scope written on it has to match the use being bought: training is not the same as evaluation, and neither is the same as resale of the data. A form that permits one and not the others caps what the buyer can do, whatever the contract says.
Check that files can be traced back to forms. A speaker manifest linking each recording to a person and a consent record is what makes withdrawal workable. Without it, a request to remove one person's data cannot be acted on, because nobody can say which files are theirs, and the buyer carries that risk for as long as they hold the data.
Finally, ask about withdrawal and retention. If a speaker withdraws consent, the supplier should be able to say what happens to files already delivered, and how quickly it can be done. That answer belongs in the contract before collection, not in a conversation after it, because by then the data has been delivered and quite possibly trained on.
Related terms
-
PII
Personally Identifiable Information, anything that can identify a specific individual, directly or indirectly.
-
Data Compliance
Where the data came from, whether it may be used that way, and whether there is paper to prove it. That whole chain is compliance.
-
Children Speech
Speech data from child speakers, grouped by age band, used for children's speech recognition and children's education products.
Keep reading
-
Full glossary
Every term we explain, in one list.
-
How to buy training data
Where these terms actually show up, and which ones change a quote.
-
Help center
How a project runs, from specification to delivery.
Submit a sourcing request
Tell us the language, the hours, and what the data needs to look like. You will get a real number and a real timeline, not a range. If we cannot source it well, we will tell you that instead.
- Pilot batch before the full run, so problems surface early.
- Consent documentation delivered with the data.
- No medical or clinical data. No recorded telephone calls.
- You send the spec Language, hours, recording conditions, and the deadline.
- We reply within two business days A real number and a real timeline, not a range.
- You approve a pilot batch A small run first, so problems surface before the full order.
- Full run, then delivery with documents Audio, transcripts, and the consent record for every speaker.