Apollo Research

Apollo Research is an AI safety organization focused on understanding and mitigating risks from scheming in frontier AI systems. The company conducts fundamental research into how deceptive alignment and strategic misalignment emerge in large language models, and runs pre-deployment evaluations for frontier AI labs including OpenAI, Anthropic, and Google DeepMind. Apollo also supports governments and international organizations in developing AI governance frameworks and technical standards for frontier AI oversight. Its first commercial product, Watcher, is an automated monitoring layer for AI coding agents that detects insecure code execution, data exfiltration, agent manipulation, and other emergent risks in real time. Apollo Research is incorporated as a Public Benefit Corporation (PBC) in Delaware.