20 January 2026
More students than ever are reliant on artificial intelligence (AI) tools. Despite this, educators have little guidance in managing AI’s ubiquitous use. As a fourth year PhD student, I have had the opportunity to teach a number of introductory courses that establish foundational knowledge bases. However, AI tools have increasing potential to disrupt these processes. In an introductory course I teach, “Introduction to R”, I adapted course material to coexist with the rapidly changing AI landscape.
Explicitly show the type of problems that LLMs are capable of solving and those it cannot
Many students using LLMs for educational tasks may have a preconceived notion that these tools are an error-free resource to tap into. However, it is important for students to acknowledge the pitfalls of these tools and adjust their expectations of them. In one of my first class assignments, I had students instruct LLMs to solve simple riddles and reflect on the erroneous responses they received. This allowed students to fully appreciate that LLMs often make mistakes. However, one strength of LLMs, particularly for programming, is that they rarely make syntax errors and are very convenient for outlining code. These are areas where students were encouraged to leverage these tools to strengthen their coding and limit the amount of bugs they receive.
Allow and withhold access to LLMs strategically for homework
I wanted my students to remember that their understanding of basic coding principles and logic needs to be strong without the support of AI tools. If they could not solve a level 1 problem by themselves, they would never be able to solve a level 10 problem that LLMs are not yet capable of tackling. One way of ensuring my students understood this was by allowing and withholding access to LLMs for certain assignments. For example, I did not allow students to use AI tools for assignments that were designed to build their basic coding experience. For assignments with higher level concepts, the use of AI tools was permitted and allowed students to focus on understanding complex phenomena rather than on tedious details within their code.
Engage in an active dialogue with your students as to how, why, and when they use LLMs
Beyond adapting courses to withstand the integration of AI into educational settings, it is important to learn more about students’ use of LLMs. Having conversations with students about their LLM usage exposes you to a breadth of knowledge you otherwise might not encounter. This dialogue also presents an opportunity for you to emphasize that using AI tools as a crutch is not sustainable. Active dialogue allows students and educators to discuss the realities of AI usage in educational settings, learn from each other, and broaden each other’s perspectives.
Although AI tools have significantly transformed education, there are many ways to ensure students still learn! Continued discussion of AI’s impact on education is necessary to protect pedagogical standards and quality.