Verification Network

Score the AI ecosystem.

182 rateable platforms. Transitioning from subjective rubric scoring (0–3 per dimension) to verifiable binary features — facts anyone can check. The existing 1,344 scores use the framework’s own rubric; feature-based scoring eliminates that circularity. Verify feature codings, challenge errors, track changes over time. Evidence-based, adversarial, bounty-driven.

1,344+ platforms scored
community submissions
features verified

How it works

1
Pick a platform

Choose from the queue of platforms that need more ratings, or nominate one that isn't scored yet.

2
Verify features

Check verifiable binary/ordinal features — does this platform have an algorithmic feed? Does it publish a model card? Provide evidence: URL, screenshot, or archived link.

3
Earn ATH

Verified features earn ATH rewards. Successful challenges earn bonus ATH. Build reputation through accuracy, not volume. Top verifiers become arbiters.

How verification works

Each feature is a verifiable fact — does this platform have an algorithmic feed? Does it publish a model card? Multiple verifiers check the same features independently.

ICC (Intraclass Correlation) measures agreement between verifiers. ICC ≥ 0.60 means strong consensus. Disagreements are flagged for arbitration.

Read the methodology →

Top contributors

Loading…