Data & consent
How we handle your clinical judgment.
You should know exactly what happens to your answers before you give them. This page explains it in plain language. The legal version is in our Privacy Policy.
What an activity produces
When you complete a clinical activity, four things are recorded:
- Your assessment — the rating or decision you committed to.
- Your reasoning — what you said was missing, wrong, or important.
- Your confidence — how sure you were.
- Your professional context — credential, specialty, subspecialty, years in practice, board status, and practice setting.
That last one is what makes the data useful. A judgment from a shoulder specialist and a judgment from a generalist are different information, and both matter.
What leaves Evident Labs, and what never does
Never shared
- Your name
- Your email address
- Your licence number
- Your NPI
- Your employer or institution
- Anything that identifies you personally
May be shared, de-identified
- Your assessment and reasoning
- Your stated confidence
- Non-identifying context, e.g. “orthopaedic surgeon, 10+ years”
- Aggregate agreement across many clinicians
Datasets we deliver contain clinical judgment, not clinicians. A partner receiving our data cannot tell that any particular answer came from you.
Why we ask for it
Medical AI is being evaluated against something. Today that something is often a benchmark written by people who don't practise, or a consensus scraped from text that may be years out of date.
We think it should be the judgment of practising clinicians, captured properly — independently, before consensus is visible, with the disagreement left in.
You are not labelling data for a model. You are setting the standard the model has to meet.
What you control
You can opt out and keep your CME.
Contributing to evaluation datasets is separate from earning credit. Turning it off costs you nothing.
You can see what you've contributed.
Your activity history is yours to review and export.
You can delete your account.
We remove your personal information, keeping only what accreditation record-keeping requires.
Opting out applies going forward.
We can stop including your judgment in future datasets. We can't retract it from a dataset already delivered, and we won't pretend otherwise.
Things we don't do
- We don't sell your personal information.
- We don't ask for patient data. Our cases are constructed or de-identified, and you shouldn't submit PHI.
- We don't average away disagreement. Where medicine genuinely splits, we record the split.
- We don't share your identity with AI partners, ever.
If you're evaluating us as a data source
The same properties that protect clinicians are what make the data worth buying.
Verified provenance
Every judgment is attributable to a credential-verified clinician. We know who answered, even though you don't.
Independent capture
Clinicians commit before seeing peer responses, so agreement is real rather than anchored.
Documented consent
Clinicians are told what their judgment is used for and can decline without losing access to the product.
Where we are
Evident Labs is early. The commitments on this page are decisions we've made, not features we're describing — they hold from the first activity we run. Where a control needs an interface, such as the opt-out setting, it will exist before accounts open.