GDPR Article 25 (“Data protection by design and by default”) is not a nice-to-have for online exams. It makes privacy-preserving defaults a legal expectation: the most privacy-friendly configuration should be the baseline, not an optional toggle.
For proctoring, the core implication is simple: collect only what is strictly necessary to protect exam integrity.
What “privacy by default” means for proctoring
A privacy-by-default proctoring setup typically implies:
- Data minimization: avoid uploading raw streams if integrity can be established with less
- Purpose limitation: signals used for integrity only, not secondary analytics
- Retention minimization: keep evidence only as long as needed for disputes and audits
- Least intrusive defaults: proctoring intensity should match exam risk level
“State of the art”: on-device processing vs cloud streaming
Traditional tools stream continuous video and audio to the cloud. That creates:
- more personal data collected by default
- greater transfer and storage risk
- stronger obligations around access control and third-party processors
In contrast, on-device analysis can process signals locally and discard non-essential data immediately—supporting the “state of the art” principle by reducing exposure while still producing integrity-relevant logs.
A practical checklist for exam owners
- Can the vendor operate without storing raw video?
- What events are logged, and can a reviewer reconstruct what happened?
- Can the monitoring level be configured per exam type (low/medium/high stakes)?
- Are defaults conservative, or do they start from maximum surveillance?