James Landay Explains Why AI Should Be Human-Centered
James Landay argues that artificial intelligence must be designed with human needs at its core to ensure a positive societal impact. This approach prioritizes the integration of AI into daily life by focusing on design choices that benefit people rather than purely technical metrics. The core premise is that widespread adoption requires systems explicitly built to serve human interests and societal well-being. Without this human-centered framework, the technology risks failing to deliver value to the populations it is intended to serve.
Course Overview: AA228 Decision Making Under Uncertainty
The course AA228 teaches students to optimize decisions when outcomes and models contain uncertainty, a challenge rooted in Richard Bellman's 1950s work on the curse of dimensionality. While provably optimal solutions are often impossible for complex problems like wildfire fighting, the curriculum focuses on developing appropriate approximations and evaluating computational methods against baselines. Students apply these statistical techniques to diverse domains, emphasizing the critical need to validate AI systems and quantify risk before real-world deployment. The ultimate goal is to design decision-making agents that deliver societal benefits while ensuring responsible behavior under limited information.