A professor at a mid-sized university informed a colleague a few months ago that she had completely stopped giving out take-home essays. She claimed that it wasn’t because the students were indolent, but rather because she could no longer identify the author. In higher education, that modest, almost embarrassing admission has become commonplace. The process of rethinking education for the AI era at universities appears to be much more complicated than any official press release would imply.
It wasn’t a sudden change. The majority of institutions responded to the rapid spread of ChatGPT and similar tools on campuses in the same way that anxious institutions typically do: they first imposed restrictions before considering other options. Some outright prohibited AI. Others issued ambiguous directives that no one could fully comprehend. Some chose to do nothing in the hopes that the moment would pass. It didn’t. Whether or not their school approved of it, surveys and faculty interviews conducted since then indicate that nearly every student in nearly every discipline has used an AI tool for coursework at some point.
The way the discussion has progressed is intriguing. Whether or not students are cheating is no longer a major concern. The question is whether the lecture, assignment, and exam formats of the traditional teaching model still make sense in light of the fact that a chatbot can generate a well-written essay in a matter of seconds. In a recent industry briefing, Danny Liu, an educational technologist at the University of Sydney, stated bluntly that institutions have been talking for more than two years without taking any concrete action. Compared to most official statements, that kind of admission from someone who is part of the system says more.
Some educators believe that the true threat is not AI per se, but rather what would happen if students were never taught to think without it. Recently, researchers from Curtin University and the University of Melbourne contended that motivation, rather than technical proficiency, is the real issue. According to their research, students were much more likely to apply what they learned if they started to see themselves as future professionals rather than as individuals pursuing grades. The students who were most inclined to let AI think for them were those who were still in “jump through the hoops” mode. The typical solution of simply adding more AI-related assignments to a syllabus is complicated by this subtle but significant difference.
Redesigning courses so that AI becomes a part of students’ reasoning rather than a shortcut around reasoning is one way that some universities have begun to organize their response into rough stages, moving from outright restriction toward something closer to integration. In reality, very few schools have attained that higher level. The majority are still in the middle, experimenting erratically and frequently motivated more by a small number of driven professors than by a well-thought-out institutional strategy. Walking through any large university building reveals this disparity: a department on one floor may be conducting pilot programs using AI-assisted feedback tools, while another department two floors up still views any use of AI as grounds for an academic integrity hearing.
Beneath all of this is a more subdued and unsettling question that a Pakistani student almost unintentionally brought up in a recent opinion piece. Was the system ever as stable as it appeared before AI arrived if schools have spent decades shaped more by enrollment figures, corporate sponsorships, and strict staffing ratios than by actual learning outcomes? It’s possible that rather than ruining higher education, AI revealed preexisting flaws.
None of this indicates that degrees are going to become less valuable or that lecture halls are going extinct. However, it’s difficult to ignore the fact that the organizations that take AI seriously aren’t the ones loudly opposing it. They are the ones posing more challenging questions about what a graduate ought to be capable of and being forthright about the distance they still need to cover.
