We work with research labs, supervisors, and industry partners who share our commitment to rigorous, evidence-based AI safety science.
- Co-supervision of students on frontier safety problems
- Joint research on interpretability, alignment, agentic safety, or evaluation
- Safety audits, shared benchmarks, and evaluation infrastructure