Validate AI Findings
Compensation System
We have designed our system in a way that promotes the accuracy of validations. Participants evaluate AI-generated issues over a one-month period with performance-based rewards.
Evaluation Process
100 AI findings are uploaded for review every month
Each finding is initially labeled by 3 validators; if they have full agreement (3/3), the validation is accepted, and the finding is labeled and removed from the evaluation. If the validators don’t agree on the validity of the finding, the finding is disputed.
Disputed findings receive 2 more evaluations. In the case of a 4/5 agreement, the finding is accepted, labeled, and removed from the evaluation. If there is no such agreement, the review is escalated to the Head of Security of AuditOne who reviews it and labels it.
Senior Auditor from AuditOne checks a random sample of findings for accuracy calculations.
Performance & Payment
Auditors are evaluated based on Senior Auditor labels. Payments are in stablecoin at month-end.
Base Pay: $2 per labeled finding.
Total Pay = Base Pay + Accuracy Bonus.
Accuracy
Bonus/Finding
≥ 90%
$1.50
80–90%
$0.70
70–80%
$0.30
< 70%
No base payment and no reward
Example: You labeled 100 findings; 17 spot-checked, 15 matched → 88% accuracy (Tier 2) → *$270 earned ($200 base + $70 bonus).
Last updated