If you've ever caught yourself in a conversation with a colleague and wondered whether you were actually talking to them or to whatever chatbot drafted their last email, you already have a feel for why AI literacy has quietly become a PETE issue. It's easy to file AI under "someone else's problem" — a computer science concern, a K-12 policy question, not something that touches lesson planning or gym space. Dr. Xiaolu Liu, lead author of a new call-to-action piece in Quest titled "Integrating Artificial Intelligence (AI) Literacy into Physical Education Teacher Education," argues that this framing is already out of date. I talked with her on the podcast about what the paper actually calls on PETE faculty to do, and the case is a compelling one.
Using AI Isn't the Same as Being AI Literate
The paper's central move is a distinction that's easy to miss: being able to type a prompt into ChatGPT does not make a pre-service teacher AI literate. Drawing on an OECD framework, Liu and her co-authors describe AI literacy as three interlocking pieces — knowledge, skills, and attitudes/ethics. Knowledge means understanding that a large language model isn't retrieving facts; it's predicting plausible-sounding text based on patterns in its training data, with no awareness of your students, your gym, or yesterday's lesson. Skills mean being able to evaluate what comes out the other end — is this rubric actually measurable? Developmentally appropriate? Aligned to the objective? And attitudes/ethics keep the responsibility where it belongs: with the teacher, not the tool. As Liu put it, the real question isn't whether a PE teacher can use ChatGPT — it's whether they can make an informed, professional judgment about when AI should be part of the decision and when it shouldn't.
Prompting as a Window Into Pedagogical Thinking
One of the more useful reframes in our conversation was around prompt engineering. Liu doesn't want faculty teaching students a "perfect prompt" formula. Instead, she treats a vague prompt versus a detailed one as a diagnostic: a strong prompt requires the student to already be thinking like a teacher — grade level, space and equipment constraints, class size, students needing modifications. A simple in-class activity she suggested: give everyone the same vague prompt, have them identify what's missing from the output, then have them revise the prompt with real teaching context and compare results. The prompting itself becomes a way of making pedagogical thinking visible — which is exactly the habit of mind PETE programs are already trying to build.
You Don't Have to Be an AI Expert to Mentor This
If you're a mid-career faculty member who has poked around with AI out of curiosity but hasn't brought it into your courses, Liu's advice was refreshingly low-stakes. You don't need computer science expertise — your expertise is pedagogy, curriculum, and assessment, and that's precisely what the AI conversation is missing. Her suggested starting point: take one task you already know well (she uses assessment development in her own courses), generate something with AI, and interrogate it with the questions you'd already ask as an expert — does this align with the objective, is it developmentally appropriate, what did the AI misunderstand. From there, talk to colleagues, compare notes, and resist the pressure to have all the answers. One of the paper's more freeing points is that faculty are allowed to be mentors and learners at the same time.
Scaffolding Without Adding a New Course
Liu was careful to say the paper isn't asking every PETE program to bolt on a new three-credit AI course — that's an unrealistic ask for an already-packed curriculum. Instead, she describes a natural progression through existing coursework: foundational courses build basic AI knowledge (what a large language model is, why it can be wrong, what bias means); pedagogical courses like assessment and methods move into application and critique, where students generate something with AI and then justify what to keep, revise, or reject; and advanced methods or field experiences push toward independent professional judgment — when AI adds value, when it interferes with learning, and what ethical concerns are in play. The ethics conversation itself — bias, FERPA and student privacy, and even the language we use about AI ("the AI generated this" rather than "the AI decided this") — isn't a bolt-on lecture but something woven through every stage.
The paper's underlying argument is one worth sitting with: physical education is relational and embodied in ways AI cannot replace, and it shouldn't be asked to. But AI literacy, done well, is precisely what protects that human core — by preparing future teachers to decide, deliberately, when a tool belongs in the room and when it doesn't.
This blog is a companion to the HPE Research Podcast episode featuring Dr. Xiaolu Liu (Georgia State University). Her paper, "Integrating Artificial Intelligence (AI) Literacy into Physical Education Teacher Education: A Call to Action," is published in Quest —
To cite this article: Xiaolu Liu, Yin-Chan Liao, Xiaoping Fan, Liyan Tang & Deborah Shapiro (29 May 2026): Integrating Artificial Intelligence (AI) Literacy into Physical Education Teacher Education: A Call to Action, Quest, DOI: 10.1080/00336297.2026.2680060
This blog post was written with the assistance of AI to support clarity and accessibility. It is intended to help disseminate and discuss research findings with a broader audience. However, for the most accurate and reliable information—including conclusions and practical applications—please refer to the original peer-reviewed publication on which this blog is based. The peer-reviewed article remains the most authoritative source.