A new Inside Higher Ed Student Voice survey of more than 1,000 students across 166 two- and four-year institutions cuts through a lot of the noise about generative AI in higher education. The headline finding? AI is changing college, but it isn’t killing it. For institutions trying to chart a thoughtful path forward, the data offers a clearer picture of what students are actually doing with AI — and what they want from their schools.
Here are the takeaways every academic leader, faculty developer, and instructional designer should be sitting with.
1. Students are using AI — but mostly to support learning, not outsource it
Eighty-five percent of students reported using generative AI for coursework in the past year. But the dominant use cases look a lot more like productive study habits than wholesale cheating: brainstorming ideas (55 percent), asking AI questions like a tutor (50 percent), and studying for exams (46 percent). Only 19 percent reported using AI to write full essays.
The implication for institutions: treating AI use as a binary “cheating vs. not cheating” issue misses the texture of how students are actually integrating these tools. The harder, more useful work is helping students discern between uses that build skill and uses that erode it.
2. Grade pressure — not confusion about policy — is the main driver of misuse
When students were asked why their peers misuse AI, the top answer was pressure to get good grades (37 percent), followed by time pressure (27 percent) and indifference to academic integrity policies (26 percent). Only 6 percent blamed unclear policies.
This reframes the conversation. The problem isn’t primarily that students don’t know the rules — it’s that the incentive structure around grades is fueling shortcut-taking. Institutions that focus only on tightening policy without addressing assessment design and grade-driven anxiety are treating a symptom, not the cause.
3. Students want action — but not surveillance
Ninety-seven percent of students believe institutions should respond to AI-era integrity challenges. But the popular interventions aren’t the ones often debated in faculty senates. AI-detection software (21 percent) and tech restrictions in classrooms (18 percent) ranked low. What students actually want:
- Education on ethical AI use (53 percent)
- Clearer, standardized policies on when and how AI can be used
- Flexibility to use AI with transparency
The lesson for institutions: a punitive posture is misaligned with what your students are asking for. Educational, transparency-based frameworks have a far stronger mandate.
4. Students are skeptical of faculty AI use, too
Only 29 percent of students felt positively about faculty using AI to create assignments or do other teaching work, while 39 percent felt negatively. The concerns students raise about faculty use — quality, overreliance, opacity — mirror the concerns faculty raise about student use. Any institutional AI strategy that asks more of students than it asks of instructors will struggle to land.
5. The critical thinking question is real — and nuanced
Among students who use AI for coursework, 55 percent said it has had mixed effects on their learning and critical thinking. Twenty-seven percent said the effect is positive; 7 percent said it’s negative. Notably, students who use AI to write essays were nearly twice as likely as those who use it to study to report negative effects on their thinking.
This is the data point worth holding onto. AI’s impact on cognition isn’t uniform — it depends heavily on how it’s used. That’s a design problem institutions can address.
6. Students want AI literacy as career preparation
Looking beyond the classroom, students said they want institutions to offer (but not require) training in professional and ethical AI use, alongside clearer guidance and open forums to discuss risks and benefits. Only 16 percent thought AI preparation should be left to individual professors, and just 5 percent thought colleges should do nothing. The expectation of an institutional response is overwhelming.
7. AI hasn’t devalued college — and for some, it’s raised the value
Thirty-five percent of students said AI hasn’t changed how they value their degree, and 23 percent said college is more valuable now. Only 18 percent said they question its value more than before. The “AI will kill college” narrative isn’t reflected in how students themselves think about their education.
What this means for institutional strategy
The survey points toward a clear playbook for institutions ready to lead rather than react:
Move from policing to pedagogy. Invest in AI literacy curricula and assignment-by-assignment guidance rather than detection tools that students distrust and that often produce false positives.
Redesign assessment. If grade pressure is driving misuse, then scaffolded assignments, alternative assessments, and a refocus on learning over point-getting do more for integrity than any policy ever will.
Build institution-wide frameworks. Students want consistency, not contradictory class-by-class rules. Frameworks like the University of Kentucky’s Student AI Use Scale or Mark Watkins’ VALUES framework offer starting points worth studying.
Prepare students for the workforce. Adult learners especially want training on how AI shows up in professional settings. This is a place where higher ed can lead.
There is, as one survey reviewer put it, “no instruction manual” for this moment. But the students themselves are telling us — clearly — what they need. The institutions willing to listen will be the ones that turn this transition into an opportunity instead of a crisis.
Based on “How AI Is Changing — Not ‘Killing’ — College,” by Colleen Flaherty, Inside Higher Ed, August 29, 2025. Read the original article.