On March 9, 2026, zero-knowledge-proof computing company Brevis unveiled Brevis Vera, an end-to-end media authenticity system designed to cryptographically verify the origins and editing history of images and videos while protecting user privacy. The launch marks a strategic shift from reactive detection to proactive proof in the fight against deepfakes.
From Detection to Proof: Vera's Paradigm Shift
With deepfake technology advancing rapidly, the industry has relied heavily on AI-based detectors to identify manipulated media. However, Brevis argues that this is a losing battle — detectors are always playing catch-up. Vera takes a fundamentally different approach: instead of asking whether media looks real, it enables media to prove where it came from. This cryptographic proof is immutable and verifiable without relying on any trusted third party.
Traditional C2PA signatures authenticate files at the moment of capture, but any edit — cropping, resizing, or compression — breaks the signature. Users face a dilemma: either use raw, unedited files (often impractical) or lose the proof of authenticity. Vera bridges this gap with Brevis Pico, a zero-knowledge virtual machine (zkVM) that generates mathematical proofs confirming that edits are legitimate transformations of the original. This preserves the authenticity chain throughout the entire workflow.
Three-Phase Workflow: From Lens to Screen
According to Brevis's blog post, Vera operates in three integrated phases:
- Capture Signing: C2PA-enabled devices sign the media at the source.
- Edit Proving: As the file undergoes edits or compression, the zkVM generates proofs that each modification is a valid transformation.
- Platform Display: End-user platforms display a "Vera-certified" badge, offering a complete, verified history of the digital asset.
This end-to-end workflow ensures authenticity from the camera lens to the viewer's screen, with no trust assumptions beyond the cryptographic proofs.
A Critical Moment: Deepfakes Threaten Global Stability
Vera arrives at a time of escalating digital distrust. In the first quarter of 2026 alone, multiple high-profile deepfake incidents have shaken the world: hyper-realistic impersonations of government officials, AI-generated videos targeting political candidates during major international elections, and sophisticated audio forgeries used in financial scams.
These events have fueled what experts call the "Liar's Dividend" — a phenomenon where the mere existence of deepfakes allows bad actors to dismiss genuine evidence of corruption, war crimes, or abuse as AI-generated. The default reaction to striking imagery has shifted from curiosity to suspicion. Brevis believes that Vera can flip this dynamic by making authenticity the default, restoring verifiability to truth.
While Brevis has not yet announced commercial pricing, it has opened Vera's developer interface for integration. The company aims to partner with news organizations, social media platforms, and content moderation services to tackle the authenticity crisis in the AI age.