What Humans Want: Models of Preference and Reward in Minds and Machines

Brian Christian gives his talk, “What Humans Want: Models of Preference and Reward in Minds and Machines,” on October 7, 2026.

EECS Colloquium

Wednesday, October 7, 2026

Gateway 1220
4:00 – 5:00 pm

Brian Christian

Research Fellow at UC Berkeley’s Center for Human-Compatible AI

Bio

Brian Christian is a Research Fellow at UC Berkeley’s Center for Human-Compatible AI, where his work spans AI alignment and computational cognitive science. His research has appeared in venues like ICLR, FAccT, COLM, and PNAS and been covered by Nature and The New York Times. He is the bestselling author of three books, including Algorithms to Live By (with Tom Griffiths) and The Alignment Problem, which won the Excellence in Science Communication Award from the National Academies of Sciences, Engineering, and Medicine. He has lectured at Google DeepMind, OpenAI, Microsoft, and Meta, and advised policymakers in six countries and the UN.

Abstract

AI alignment rests on a simple model of human decision-making: people are noisily rational agents maximizing fixed rewards that exist in the environment. This talk examines that model from both sides: minds and machines. In humans, choices follow resource-rational heuristics rather than optimal policies, and pursuing even an arbitrarily assigned goal devalues the alternatives. In machines, reward models trained on human preferences show striking biases, many inherited from pretraining rather than from the preferences themselves. Finally, LLMs share humans’ inability to “un-know” biasing information, but, unlike humans, can converse with a blind, counterfactual self, substantially mitigating demographic bias and sycophancy.