Ethics in Computing

We’re working to scale safe oversight of AI through mechanistic interpretability.

Our mission is to understand how AI systems work internally and turn that understanding into practical tools for oversight, so human understanding and control can keep pace with AI capabilities.