Skip to content
HR privacy9 min read

How to redact faces in HR recordings without creating another data copy

A local workflow for preparing interviews, training footage and workplace recordings for controlled disclosure.

Practical outcome: Minimise unnecessary disclosure while keeping a clear human approval step.

01

Start with purpose and audience

Before editing, identify why the recording is being shared and who needs access. Redaction should support data minimisation, not replace an access decision.

Create a working copy in an approved location. Keep the original unchanged with narrower permissions and record the relationship between the two files.

02

Review more than the primary speaker

Workplace video may reveal colleagues in the background, names on screens, badges, documents and reflections. Face censorship addresses one category of identifier; the reviewer must still inspect the complete frame and audio.

  • Background employees and visitors.
  • Mirrors, glass and monitor reflections.
  • Name badges, whiteboards and screen content.
  • Spoken names and sensitive conversation.

03

Approve and retain deliberately

Use a second-person check for sensitive disclosures. The reviewer should verify mask continuity, false positives and the final output destination.

Set retention for temporary renders and review reports. Local processing reduces transfers, but good file handling remains essential.

Apply the workflow

Review real footage locally.

CensorFlow detects and tracks visible faces on your Windows PC, then lets a person correct and approve the result before export.