Ask
Ask a question and get an answer drawn only from published evidence.
Ask the Council
Ask about an AI model's scores, or paste a signed record to check it. Answers come only from published measurements — if there is no evidence, the Council says so.
- See the scoresHow each measured AI model did on every published test.
- Check a signed recordRecompute its fingerprint and signature in your own browser. Free.
- Replay a model arenaRecorded rounds between models, graded by fixed rules.
- Learn how the tests workWalk through a test step by step, then try it yourself.
Connect your own AI tool instead? See the 16 MCP tools
Account overview and recent measurements
RLMAI Recommendations
AI-powered compliance recommendations that learn from your interactions
Critical
0
High Priority
0
Medium Priority
0
Low Priority
0
No Recommendations
Register AI systems and run assessments to receive personalized recommendations.
How RLMAI Recommendations Work
Incident Analysis
Analyzes Watchdog incident patterns to identify emerging risks.
Scenario Learning
Uses scenario inputs; no live 33-seat voting record exists.
Personalized Actions
Tailored to your AI systems and compliance gaps.
Continuous Learning
Improves based on your feedback and interactions.
Answers come only from published measurements; nothing you paste is uploaded.