Trust and privacy
How SCIO keeps AI under human control and protects records.
Human review comes first
Every AI-generated question is independently verified and instructor-approved before any student sees it.
-
Generated
Questions are drafted from the module's content, topics and scoped knowledge sources.
-
Independently verified
A separate correctness check marks each question.
- Agreed
- Disagreed
- Inconsistent
-
Instructor-approved
You approve, reject or edit each question. Nothing reaches a student unreviewed.
What the verification marks mean
Each generated question receives one of three marks. They help you decide where to look first.
Agreed
The independent check agrees with the question. You still review and approve it.
Disagreed
The check does not agree. The question is flagged for your attention and left out of bulk approval.
Inconsistent
The check gave mixed results. The question is flagged so you can look at it closely before deciding.
Records, isolation and data protection
What SCIO does to keep records accurate and personal data protected.
Immutable audit and override records
Every override needs a reason and is recorded in a trail that cannot be edited afterwards.
Organization isolation
Users only see their own Organization's data.
Export and anonymization
Export a student's full record set. Anonymize on request: a reason is required and recorded, the change is irreversible, and de-identified data is kept for statistics.
Encumbered-content warnings
Licensed sources are flagged, so you know before you build on them.
Fail-closed personal-data check
Generation requests that still contain personal data are blocked. Snapshots are redacted, never raw.
Invite-only access
Instructors join by single-use, expiring email invite. There is no open sign-up.
For how we handle personal information on this website, read our privacy policy.
Try SCIO with your own course material
SCIO is opening up gradually. Join the waitlist and we'll let you know when it's your turn.