A worried professor sent an all-caps email that captured a tension simmering across college campuses: “Can we really not detect AI use? What counts as plagiarism?”
Two years since generative artificial intelligence hit the academic mainstream, it’s clear that student use is widespread, but faculty guidelines are still inconsistent and scattered. According to one of the biggest surveys in the country focused on AI and higher education, only 14 percent of instructors report feeling confident using these new tools in their teaching.
Students notice the gap. At University of California, Davis, a survey found 70 percent of undergraduates want more instruction around AI, yet just a third of their professors have brought such tools into their courses.
Everywhere, the pattern repeats. Forty to sixty percent of students say they turn to AI weekly for help with writing, but many confess uncertainty about whether they’re using it the right way. One student summed up the mood, saying, “These tools are around and here to stay, so instructors may as well get on board and help us determine how to use them in a reliable and appropriate way.”
Some curriculum designers fear that relying too much on AI could erode essential skills. There is research to back their anxiety. One study showed that pilots overly reliant on autopilot struggle in emergencies and that GPS use has led people to lose their innate sense of direction. Recent results from high school math students in Turkey confirmed that AI can help with practice problems, but in a test without digital help, students in the unrestricted AI group actually performed worse than their peers who had only textbook support.
Balancing Risk With Reality in Higher Education
All this suggests that AI, while powerful, mirrors other technologies that boost short-term performance at the expense of lasting skills—unless students get strong guidance on how to use it wisely.
Flat-out bans are often rejected because enforcing them is nearly impossible right now, and pushing the technology into the shadows only drives creative misuse. The Turkish data were clear: using AI as a tutor is more beneficial than deploying it solely as a shortcut. The challenge now is figuring out which skills remain essential, especially in disciplines like scientific writing, as academic policy on AI use changes the jobs students will do in the future.
For some professors, like those teaching graduate-level writing, the solution is carefully calibrated. Students can use AI to fix grammar, clarify structure, and double-check they have not missed items on a checklist. They are prohibited from using AI to draft research ideas or build experiments, because those are the deep thinking skills that must be developed through individual effort.
Motivation reinforces all of this. When a student asks, “Why not just let AI do all the work?” the analogy is simple: lifting a barbell yourself builds muscle, but if you let a machine do the heavy lifting, those muscles never develop.
The pressure is real, and not every institution will have the same resources to address these challenges. As one professor put it, “At other institutions, and with other students, generative AI technologies are wreaking havoc, as students can now graduate from college essentially illiterate.”
Shaping policies for AI use in education will require regular reassessment. Teachers and administrators alike must come together to redefine curriculum and ensure genuine learning still happens, even as the boundaries of student assignments and AI adoption continue to move.
