I am currently involved with two courses on a regular basis:
CS224M Multiagent Systems. This course covers various aspects of extending AI theories and techniques from the single-agent case to the multi-agent case. Emphasis is on foundational concepts, representation techniques and algorithms, rather than applications. Topics include modal logics of knowledge and belief, logics of belief change, multiagent probability systems, distributed problem solving, introduction to noncooperative game theory, mechanism design and auctions, cooperative and non-cooperative communication, and multiagent learning.
CS222 Knowledge Representation. I co-teach this one with professor Johan van Benthem from Philosophy, and the course is cross listed as Phil 358. The philosophy course is titled "Rational Agency and Intelligent Interaction", which is a more apt title for the class. It looks at the use of modal, temporal and dynamic logics to reason about knowledge, belief, their dynamics, agency and intention, and some aspects of game theory. See course website for details.
From time to time I also teach a Freshmen Seminar. This year the seminar, CS21N, is titled "Can Computers Think? Can They Feel?" Previously I taught one titled "The Role of Information Technology in Global Conflict Resolution."
Other courses I have taught in the past are CS221/121 (Introduction to Artificial Intelligence), CS227 (Algorithmic Techniques in AI), and CS206 (Technical Foundations of Electronic Commerce).