By Jennifer Zaylea, Assistant Director of Emerging and Digital Technologies
Throughout the teaching profession, we keep hearing that “the AI can do it.” Some instructors are even experiencing a sense of existential dread at the prospect that they no longer need to teach our students certain skills—in some cases, skills previously thought to be crucial to their field.
But is that really true? Is generative AI (GAI) perfect? Does its knowledge exceed the human knowledge that created it? These questions are stumbling blocks for faculty walking into class every day. Students often don’t see how GAI might be undermining their learning, and some faculty are beginning to wonder whether foundational skills still matter at all.
Consider a simple example. A GAI tool can write code for my project, but that doesn’t mean the code will produce the output I want across multiple runs. I still have to verify it against the demands of the project. If I can’t understand a line of code, I can’t prompt the tool in a useful way to identify and resolve an error, and I can’t responsibly move forward. Without foundational knowledge, we cannot examine GAI output in any deep or meaningful way. Instead, we become subjects of automation bias, the tendency to trust automated output over our own judgment.
An Old Problem with a New Face
The idea that automation diminishes human capability is not a new complaint. It is a well-established principle of human-system interaction. In her seminal 1983 paper, “Ironies of Automation,” cognitive psychologist Lisanne Bainbridge described the paradox: the more advanced an automated system becomes, the more crucial the human operator’s skills become. When the system encounters an anomaly or fails, a human must step in. But if that person has not been practicing manually then their skills have eroded and they no longer have the expertise needed to fix the problem.
Engineer and former Chief Scientist of the U.S. Air Force Mica R. Endsley extends this argument to AI in her article “Ironies of Artificial Intelligence.” In higher education, we see the same pattern of automation bias but frame it as deskilling. Skimming or lightly editing an AI-generated paragraph may feel efficient, but it bypasses the cognitive effort that drives deep, transferable understanding (Bauer et al., 2025). Over time, this weakens the development of disciplinary knowledge and makes it harder for students to recognize when an AI output is biased, superficial, or simply wrong.
Teaching Foundations in the Age of GAI
None of this means students should be kept away from GAI. It means we, as faculty, still need to instill the foundational knowledge of our disciplines. The question of the hour is how to do that without encouraging over-reliance on the tools.
One option is to limit some assignments to in-person, handwritten work. Used exclusively, though, this approach eventually becomes a disservice to students, who also need to learn how to evaluate both the input and the output of GAI. A more balanced strategy is to teach GAI fluency directly, using the tool as a thinking partner or tutor. This can reinforce the foundational knowledge you present in class by prompting students to connect concepts to their lived experiences.
You might also have students work backwards from a finished product, such as a written draft of a literature response or a completed coding project. Ask them to dig into what makes up the final piece by analyzing its underlying elements, like citations and code syntax. Teach them to verify output as they work through the process. Finally, help them identify moments when they accepted an incorrect GAI output and understand how they missed it. (Hint: they probably lacked the foundational knowledge to catch the error.)
Talking with Students About Why It Matters
Understanding why foundational knowledge matters is only half the challenge. Faculty also need language that resonates with students in the moment. When students question the value of a skill GAI can perform, the goal is not to dismiss the concern. It is to help them see the difference between completing a task and developing the capacity to do meaningful work.
Students may argue that it takes too long to learn the foundations, or protest, “But I’ll never use this skill again. Why should I have to learn it?” That reaction is understandable, but the need remains, and productive struggle is where real learning happens. The goal is often not mastery of one specific task but the development of ways of thinking that transfer to new situations. Problem solving, analytical reasoning, evidence evaluation, and communication stay valuable throughout life, and broad, well-developed knowledge structures often pay off in unexpected ways.
According to How People Learn II, expertise develops when learners build interconnected knowledge structures that help them recognize patterns, organize information, and reason through unfamiliar challenges. Experts are distinguished not simply by what they know, but by how they connect and apply it. So try reframing the conversation. Move away from “Will I perform this exact task in the so-called ‘real world’?” and toward a question the research supports: “What capabilities is this helping me develop that will transfer to new situations?”
References
Bainbridge, L. (1983). Ironies of automation. Automatica, 19(6), 775–779. https://doi.org/10.1016/0005-1098(83)90046-8
Bauer, E., Greiff, S., Graesser, A. C., Scheiter, K., & Sailer, M. (2025). Looking beyond the hype: Understanding the effects of AI on learning. Educational Psychology Review, 37(2), 45. https://doi.org/10.1007/s10648-025-10020-8
Endsley, M. R. (2023). Ironies of artificial intelligence. Ergonomics, 66(11), 1656–1668. https://doi.org/10.1080/00140139.2023.2243404
National Academies of Sciences, Engineering, and Medicine. (2018). Knowledge and reasoning. In How people learn II: Learners, contexts, and cultures (Chapter 5). The National Academies Press. https://doi.org/10.17226/24783
Jennifer Zaylea, MFA, serves as Assistant Director of Emerging and Digital Technologies at Temple University’s Center for the Advancement of Teaching.