How does AI impact Introductory Psychology teaching – a case study by peer mentors and a professor
Katrina Wang UA’27, Mäusi Ruhland-Mauhs UA’26, and Jim Stellar
We three have been teaching an Introductory Psychology course this spring 2026 term when Artificial intelligence (AI) has really kicked in. Aside from the problem of how to stop cheating on essay tests, which is important but we will not discuss here, we want to focus on how we can productively and appropriately use AI. This blog is unique because it is written by a professor teaching this course and two undergraduates who took the course before and function this year as peer mentors to try to better interface between the students and the course structure (and its much older professor).
I (KW) am a junior majoring in Chinese Studies and Accounting. I work with both educators through organizing course materials, and also undergraduate students through my role as a Resident Assistant and Peer Mentor. I remember Large Language Models or LLMs taking off during highschool- during covid, it came into prominence as students turned to online resources to aid in coursework. My high school teachers would encourage the usage of tools such as Grammarly to improve writing- something that sets off AI detection tools such as ZeroGPT. I never personally understood the reliance and usage of AI until my junior year- in the past I actively tried to avoid it, as it felt like I could grasp the concepts if I just focused and studied the materials provided. As my courseload got more advanced and specialized, I started turning to AI, for individualized practice, to learn concepts, to have step by step walkthroughs through problem sets because I couldn’t find the in person resources to help me when I needed them. The discussion of AI in education is as much a discussion on convenience as everything else.
I (MRM) am in a unique position with this issue, as I began my college career in 2021 before the advent of AI. In 2023, is when ChatGPT took off, and as students, we were first exploring its potential. Right now, it seems as though we’ve fallen into a trapdoor with this technology. From what I observe today, more students use ChatGPT than don’t. My concern isn’t so much the cheating aspect or the offloading of work, but rather our mental dependence on AI and the breakdown of our thinking capacities. We’ve collectively convinced ourselves that whatever we can produce with our minds as students can be done better by the AI. And with every assignment you offload onto chatbot it seems true! But every time we reach for technology now, instead of our own thinking machine, we reinforce our dependence on the tool. I believe this is where the real detriment lies: the loss of self-efficacy.
I find that the other side of the coin is that if prompted, it can be used as a studying tool. Over the last couple of semesters, I have used it for this purpose on a broad range of subjects. I can feed it important sources from class and ask it to extract the most important information. I can feed it a study guide from class and ask it to keep generating more practice problems. Sometimes I need a full-on back-and-forth conversation to grasp an abstract concept.
I (JS) have taught this course for years with all essay tests, a Monday quiz based on short videos to introduce the topic, a Tuesday in-person lecture on the content (e.g. learning), and a Thursday Zoom meeting with a quick review and break-out groups to get some active discussion going of the topics (e.g the importance of prediction in Pavlov’s dogs learning to salivate to a bell). The whole idea was to focus on concepts. So the essay tests are on-line, open book/notes, and feedback commented by me and the TA when graded. Our hope is that students will remember the concepts (e.g. the role of prediction in learning) and then look up or recreate from the concept of what are the exact Pavlovian stimuli and responses (e.g. US, CS, UCR, CR – look it up). Also, as an instructor, I am opposed to giving in-class blue-book essay tests as it seems to be like being a dinosaur in the modern era.
Clearly in the course the TA and instructor have to try to suppress cheating (e.g. using ZeroGPT detection, inadequate as it is). Again let’s set that issue aside for now and return to the concept learning focus. Are my students using AI as a consultant (not cut-and-paste)? Is that OK? Is it helping them to learn or is it undermining the hard work needed to learn the concepts as discussed above? In class we use an analogy that using AI is OK if it is “riding in the side-car of your motorcycle.” But do not let it “drive.”
There is a lot of talk these days of cognitive offloading vs cognitive surrender, again as discussed above. But what about cognitive mentoring? I (JS) am working on some writing with another professor at a different university. My task there was to start with the sage-on-the-state idea of the role of the professor, tracing it from antiquity, and then bringing it into the modern world of guide-on-the-side thinking. So, I decided to try AI to generate those two paragraphs like I was a student in my class answering an essay question. It was fast and a bit loopy but it had an interesting structure and good ideas especially after a few more prompts. I copied the AI-generated text into a Word Document and rewrote it. I got rid of the silliness, put it in my voice, added a few ideas of my own, and thought it was pretty good. When I ran it through ZeroGPT, I got an assessment of less than 20% AI-written. Had I gotten that assessment from an essay in one of my exams, I would not have worried about it. My experience with AI was fun. It did some good work and helped me. I told my professor colleague what I did and she said even knowing what I did that the piece was good and belonged in our writing.
Now here is the weird part. I cannot help but think that my student essays are getting better as I have taught the course over the last few years as AI has really developed. How do I get them to have my experience of learning and owning it? KW and MRM just did a survey of the course that involved AI. We will write more later but briefly now but they say that in a recent survey which did have a low (15%) response rate, 50% of the students said they only used AI “here and there” but 60% said they did not use it in this class but 70% other students did (interesting). So, we have much thinking to do about AI in the classroom, particularly where the class emphasizes concepts and uses on-line essay tests.
