Apollo Research is a leading AI Safety org. Our goal is to understand and evaluate scheming and to prevent catastrophic harms that scheming AI might cause. We build tools to make the deployment of powerful AI systems safer.