Nikhil Singh

简介

I direct the SAHAS Lab (Science and Art of Human-AI Systems), where we study the first-mile and last-mile problems of AI systems (e.g. agents) in human environments. The first mile is about understanding AI systems themselves. We draw on techniques from behavioral evaluation, interpretability (e.g. mechanistic), and representation learning to build better tools for reasoning about these systems. The last mile is about deployment: how AI systems behave in real workflows, under real incentives, and in open-ended human settings. Our aim is to make these systems more robust and beneficial under the pressures of actual use. We treat this as both a scientific and a design problem. Scientifically, we want to understand how human-AI systems work. From a design perspective, we want to shape how the