Physics Colloquium: "Teaching Students How to Learn in an AI World"

About

Please join the Department of Physics in welcoming Aaron Titus, Teaching Professor in the Department of Physics at North Carolina State University, for his talk titled, "Teaching Students How to Learn in an AI World."

Aaron Titus is a Teaching Professor in the Department of Physics at North Carolina State University where he co-coordinates the two-semester sequence of calculus-based physics. In 1997, Aaron Titus and Larry Martin created WebAssign. Titus was a member of the Open Source Physics (OSP) team, led by Wolfgang Christian, that received the American Physical Society’s 2020 Excellence in Physics Education award. Titus is a collaborator on the development of VPython and Trinket and a co-author of the fifth edition of the textbook Matter and Interactions with Ruth Chabay, Bruce Sherwood and Steve Spicklemire. In 2026, Titus received the David Halliday and Robert Resnick Award for Excellence in Undergraduate Physics Teaching from AAPT.

Abstract: Here's a fairly common scenario: students consistently attend lecture, take detailed notes, attempt homework, get stuck on homework and use AI to get the answers. They think they "understand it" when they read the solution or watch lectures, and they are confident. Then, right before a test, they spend many hours reviewing notes and practicing solving problems, perhaps asking AI for additional problems to solve. When they fail the test, they are surprised because they "spent hours studying," and their methods mostly worked in high school. In order to reach this group of students, I now give a 1-hr, optional workshop on "how to learn" for students in our large, calculus-based physics course for engineers (1700+ students per semester). The key is metacognition -- thinking about your thinking. Across three semesters and more than 3,000 students in mechanics, students who attended the workshop had a 9-point higher gain from Test 1 to Test 2 than non-workshop students, and their higher gain persisted on later tests and the final exam. Furthermore, the workshop helped change students' initial perception of the course as a "weed out" course. While the context of my work is the introductory course, the principles are applicable to intermediate and advanced physics courses where students are also tempted to relegate "thinking tasks" to AI. Metacognition is the key to using AI effectively while avoiding cognitive surrender.