Stanford Seminar - Human-AI Interaction Under Societal Disagreement

This course explores the concept of Human-AI interaction under societal disagreement. The learning outcomes include understanding how machine learning algorithms should emulate voices and labels, as well as how to use Jury Learning and The Disagreement Deconvolution to evaluate models under societal disagreement. The course teaches skills in interactive AI architecture, defining voices to emulate, and evaluating models under disagreement. The teaching method involves a seminar-style talk by a computer science PhD student. The intended audience includes developers, practitioners, and researchers interested in human-computer interaction and machine learning.