VeriFact Block Scoring
A transparent breakdown of how this video was evaluated for authenticity, integrity, and source reliability. Signals suggest the content is a candidate for verification.
Video source: SVN Submission #svn-2026-000184 · Simulated for demonstration
Composite index of source integrity, AI detection results, editorial review, and chain-of-custody signals. Scores are public, permanent, and linked to the journalist's reputational record.
Score Breakdown
Each component is evaluated independently. Scores reflect signal strength — not absolute claims about ground truth.
Uploader history, identity signals, and publishing behavior support confidence in the source.
Independent reporting, additional footage, or trusted source comparison supports the core claims.
Timestamp, location clues, weather, and known events align with the claimed time and place.
No significant signs of synthetic video, deepfake manipulation, voice cloning, or deceptive AI alteration were detected.
Frame-level review, compression patterns, and audio-visual consistency show no major signs of tampering.
An editor reviewed the technical findings, context, and edge cases before final publication.
Average across the six VeriFact layers. VeriFact performs the verification; the blockchain preserves this record as the permanent audit trail.
Reviewed by SVN editorial. A credentialed editor has evaluated AI findings alongside source context and corroborating evidence. Reviewer identity is redacted to protect editorial independence.
Verification Block Record
The public portion of this record is immutable once written. Private metadata is hashed and stored off-chain. Signals suggest this record is a candidate for verification on the VeriFact ledger.
How VeriFact Works
Every video submitted to SVN is processed through a multi-stage AI detection layer before a human editor evaluates a single frame. Machine learning models trained on millions of labeled examples analyze pixel integrity, compression signatures, frame-rate anomalies, and facial geometry — the primary indicators of synthetic or manipulated content. The result is a probability profile, not a verdict. Signals suggest the presence or absence of manipulation — not certainty.
AI analysis surfaces signals. Journalists and editors make decisions. Every piece of content that receives a VeriFact score has been reviewed by a credentialed human editor who evaluates the AI findings alongside editorial judgment, source context, and corroborating evidence. The score reflects that combined assessment — not the output of any single system.
When a video passes SVN's verification threshold, its metadata — including submission timestamp, source attribution, verification chain, and block score — is cryptographically anchored to an immutable ledger. This means the record of verification cannot be retroactively altered. If new information surfaces that changes the assessment, a new block is created and linked, preserving the full audit trail.
Transparency is not a marketing position at SVN — it is an operational requirement. VeriFact scores are public, permanent, and attached to the journalist's reputational record. Every signal used in the scoring process is disclosed. Uncertainty is stated plainly. The goal is not to claim certainty where none exists — it is to show exactly how much confidence the evidence supports, and why.