Research

Advancing the frontier of AI safety through rigorous research, open-source tools, and practical deployments.

Red-TeamingMarch 2026

6,000 Hours of Red-Teaming: How Decentralized Networks Can Solve AI Safety's Hardest Problem

A distributed adversarial research network recruited 392 independent red-teamers and produced 200 structured alignment audit scenarios over 10 weeks, contributing a large open dataset for Petri-based AI safety testing.

Red-TeamingDecentralized AIAlignmentPetri

Read Article

AI SafetyNovember 2025

AlphaPetri: Automating LLM Safety Testing with AlphaEvolve-Style Evolutionary RL on Petri

A proposal and pilot study for automating LLM safety testing using evolutionary search to generate seed prompts, addressing bottlenecks in human-crafted seed prompts for Petri's testing harness.

AI SafetyLLM TestingEvolutionary RLPetri

Read Article

Stay Updated

Subscribe to our research newsletter to receive updates on new publications and findings.