COSC 189.34 AI Agents

AI agents, defined here as software systems that perceive, reason, and act on behalf of users, are quickly transforming how we interact with computers in the real world. This course explores the design, theory, and practice of AI agents through both historical and contemporary lenses. Students will investigate how agents serve as intermediaries between human intent and computational action, with a focus on the tension between delegation and control. Through reading and discussing seminal papers spanning classical software agents, reinforcement learning agents, economic models of agents, and human-computer interaction applications, the course addresses questions like: When should we delegate tasks to agents versus maintaining direct control? How do we design agents that align with human values and goals? What can different models of delegation teach us about agent design? How do we evaluate agent behavior in complex, dynamic environments?

Course activities will include reading and discussing papers, student-led presentations, in-class debates, problem sets, and a final research project. Students will engage with both the technical foundations and human-centered implications of agent systems. The goal of this course is to prepare students to understand, critique, and design AI agents that meaningfully extend human capabilities while respecting human autonomy.