How UniFi Protect 7.2 protects against AI-altered CCTV footage

Following our recent note on UniFi Protect 7.2, we wanted to highlight one of the most important advances for anyone relying on video as evidence: the new Evidence System.
AI tools can now generate highly convincing alterations to video — changing faces, objects, timestamps, or entire events. Once footage leaves the original recording system, it can be difficult to prove what is authentic.

Protect 7.2 addresses this directly at the source:

• Every recording can be cryptographically fingerprinted with SHA-256 at the moment of capture.
• The footage is digitally signed, creating a verifiable chain of trust from the camera through to download.
• Anyone can check a clip at evidence.ui.com. Genuine video is confirmed; any alteration — including AI-generated edits — is automatically flagged.

This means the original CCTV recording remains protected. If someone tries to modify it after the fact (whether by conventional editing or generative AI), the cryptographic signature breaks and the tampering is immediately detectable.

For organisations that use video in investigations, insurance claims, or legal proceedings, this provides a practical way to demonstrate that the footage has not been altered.

If you’d like a short demonstration of the verification process or help enabling the Evidence System on your sites, just let us know.