Best of: How do you build a better robot? By understanding people.

Effective human-robot interaction requires robots to move beyond rigid, optimal models and instead account for human "bounded rationality"—the tendency for people to make decisions based on limited time, attention, and resources. Dorsa Sadigh, a professor of computer science and electrical engineering at Stanford, emphasizes that roboticists must integrate computational models of human behavior into their planning loops to navigate complex environments like traffic or healthcare settings. Rather than relying solely on data-driven approaches, which often fail to capture rare or out-of-distribution events, successful systems must anticipate human quirks and adapt to non-stationary behaviors. By developing low-dimensional representations of human intent, robots can better coordinate with people in tasks ranging from collaborative table-moving to complex assistive feeding, ensuring safety and mutual understanding in shared spaces.