Building safer AI together
AI Safety Marketplace
A searchable database connecting researchers with tractable problems in AI alignment and safety.
2
Problems
2
Open
0
Bounties
How It Works
AI safety challenges are scattered across labs, papers, and bounty programs. We're building the neutral catalog that connects problems with solvers.
1
Discover
Browse curated AI safety challenges with clear difficulty ratings, verification methods, and compute requirements.
2
Solve
Submit your research, code, or eval results with reproduction steps. Track progress and build on prior work.
3
Verify
Solutions are verified through automatic evals, human review, or research output assessment. Accepted work earns recognition.
Who It's For
Problem Posters
Post challenges you lack bandwidth to solve internally.
- AI labs and research teams
- Safety organizations
- Enterprises with AI governance needs
Solvers
Find tractable problems and contribute to AI safety.
- Independent researchers
- Graduate students
- Career transitioners and red teamers