Public Beta Now Live

SOMOS Civic Lab

Democratizing AI governance through structured public participation in red teaming exercises

Active Exercises

2,847

Participants

18,392

Flags Submitted

94%

Accuracy Rate

Why Participate in AI Red Teaming?

Join a community of civic-minded individuals helping to identify and mitigate AI risks before they impact society

Blind Model Testing

Test AI models without knowing their identity to ensure unbiased, objective safety assessments

Community Participation

Collaborate with diverse participants to surface risks that automated testing might miss

Identify AI Risks

Flag harmful content, misinformation, bias, and other safety concerns through structured exercises

Comprehensive Analysis

Contribute to detailed reports that help improve AI safety standards and governance policies

Participate in structured red teaming exercises designed by experts

Browse active exercises covering topics like election integrity, bias detection, climate information, and public services

Interact with anonymized AI models and follow structured testing guidelines to identify potential issues

Report harmful content, misinformation, bias, or other concerns with detailed annotations and severity ratings

Your findings contribute to comprehensive reports that inform AI safety standards and policy decisions

Join thousands of participants helping to build trustworthy AI systems through civic engagement