Will Fleeson

Context

What do you teach and how have you been thinking about artificial intelligence in the context of those courses?

I mainly teach classes in personality. In one of my courses, I have a rotating topic to combine with personality, and this year I’ve made that topic be Artificial Intelligence. It’s very exciting to share the discovery of this new frontier with the students. Naturally, there is a lot of uncertainty, and a lot of danger as well. This combination can be troubling.

Strategies

Has the development of generative AI changed how you teach?

Change #1: The class I taught focused itself on the relationship between personality and AI. We spent the second month of class reading original articles about personality and AI, and then in the third month, students designed studies, conducted them, and presented the results to the class, with each study concerning the connection between personality and AI. Topics included the personality types that are most likely to use AI to cheat, people’s comfort level with answering counseling-intake questions from an AI, and the different personalities that different chatbots have. 

After the students had some comfort with AI, we spent 20+ minutes generating the class policy on AI. AI first generated a list of ways AI could help students improve their work. We then went through the list systematically to discuss for each one the degree to which it should be allowed in the class. We discussed the impact different decisions would have on student learning. After that, I incorporated their input to create the class policy on AI.

Change #2: Early on, each student signed on to a chatbot and had a conversation with AI. This was to make sure they all had some experience with actual AI. As a group, we then shared our conversations. Some students planned the evening’s meal, some students had it write mock letters to ex partners, some students had it plan trips. We discussed the quality of the AI work, and many students thought the quality was not great (this was partly because they mostly were using 3.5). We then went deeper with the mock letter by trying to write prompts that brought out better letters, and we had some success with this, so there became more disagreement about the quality of the letter.

Impact

How have students responded to these changes? What have you learned and how will it shape your teaching moving forward?

One big question in the class was whether AI would be able to assist in the delivery of mental health services. Most students started the class strongly opposed to the idea. During one class, we had AI conduct a mock therapy session with myself pretending to have borderline personality disorder (this required some light “jailbreaking” to get the AI to understand I’m not actually in a crisis situation). We were all amazed at the quality of the AI therapy. By the end of the semester, most students were convinced that AI could play an important assistive role in mental health services.

Lessons Learned

What advice do you have for others who are asking similar questions and considering similar changes?

Anybody who wants to talk about AI should schedule a meeting with me! Because AI is new and unclear, the most progress might be made with a playful and flexible approach. Try random things with it and see if it can help. Ask it itself how it can help. Don’t judge it overall, but rather task by task. Try different prompts until you get some that work well.

Unfortunately, 4.0 does provide substantially better responses than 3.5 does. I would like to share some tools I’ve created with my students to assist their learning, but unfortunately, the tools require 4.0, and only some students have access to 4.0.

In the classroom, the advice I’d offer is the same that applies to other topics. Listen to the students to hear their reactions, expectations, and goals regarding AI. Work with those reactions when incorporating AI into the class.

Disciplinary Insights

How might your discipline help students prepare for a world in which AI is ubiquitous?

The interface of AI with people is going to be a highly psychologically charged one. How does AI act? How do people react? How does AI explain? How do AI and human goals align with each other? These are the kinds of questions we explored in my personality class. For example, we found that AI’s have unique “personalities”, and that the AI “personalities” affect reactions of people to AI in complex ways depending on the human’s personality.

Psychologists have been trying for decades to figure out the nature of people’s thinking, the types of goals people have, and how thinking and goals impact the quality of their lives. Now we are suddenly faced with a new “thinking” and “motivated” creature to try to understand, and that we understand surprisingly poorly. Psychologists might be able to use our knowledge and expertise to help with this.