Is AI Making College Students Dumber?
Source: Silicon Bay Partners’ Staff with assistance from ChatGPT
Photo: tcea.org
Yes—but the evidence is more nuanced than “AI is making college students dumber.” The stronger conclusion is that AI can either amplify learning or replace the mental work that produces learning, depending on how students use it.
The worrying part
A 2026 Cornell study analyzed responses from more than 95,000 students at 20 U.S. public research universities. About one-third said they regularly used generative AI to produce text, video or code for assignments, and 9% said they had used it to cheat.
And there is evidence that students themselves recognize a potential trade-off. An Inside Higher Ed survey found that 85% of students had used generative AI for coursework. Nineteen percent reported using it to write full essays, while students who used AI to write essays were twice as likely as those who used it for studying to report that AI had negatively affected their critical thinking.
There’s an obvious mechanism here:
If AI helps you understand something → you’re learning.
If AI does the thinking, researching, organizing and writing for you → you’re getting the assignment done without necessarily acquiring the skill the assignment was designed to teach.
That’s the real educational concern.
But it would be wrong to say research has established that AI overall makes students less intelligent. Several 2025–26 meta-analyses actually found positive effects on academic achievement and learning outcomes when AI is used as a learning tool. One 2026 meta-analysis covering 66 studies found a large positive overall effect of ChatGPT applications on undergraduate learning outcomes.
So how students use AI appears to matter enormously.
And yes, professors are catching students
This part is fascinating.
The latest survey of 3,000+ U.S. college faculty by the College Board found that 74% said their students were using AI to write essays or papers, and nearly half believed at least half of their students were using AI for writing-related tasks.
Harvard provides an even more striking example.
In a September 2026 survey of Harvard faculty, nearly 90% said they had received student coursework they knew or believed had been produced using AI. Yet only about 12% had referred students to Harvard’s disciplinary authorities for unauthorized AI use.
Why the enormous gap?
Because suspecting AI use isn’t the same thing as proving it.
Professors may notice:
A student’s writing suddenly becomes dramatically more sophisticated.
The vocabulary doesn’t resemble the student’s previous work.
The paper contains strange or nonexistent citations.
The student can’t explain arguments contained in his or her own paper.
A paper sounds suspiciously generic or polished.
The student’s drafts don’t correspond to the finished product.
The University of Southern California, for example, specifically recommends comparing suspicious papers with previous student work, examining the student’s writing process and checking citations and sources.
The great AI-detector problem
You might think professors could simply run the paper through an AI detector.
Not so fast.
Universities including the University of Florida, Caltech and the University of Minnesota warn faculty that AI detectors can produce both false positives and false negatives and shouldn’t be treated as proof of misconduct.
UF researchers reported in 2026 that commercial detectors produced false-positive rates ranging from 0.05% to 68.6% and false-negative rates from 0.3% to 99.6%, depending on the tool and circumstances. They concluded that the systems aren’t sufficiently reliable for high-stakes academic decisions.
That’s why a professor may have a pretty good idea that “this kid didn’t write this” but still have difficulty proving it in an academic-integrity proceeding.
Here’s the irony
AI may actually make writing assignments less useful as a way of measuring whether someone can write.
A student can potentially ask ChatGPT to:
research the subject → develop a thesis → create an outline → write the paper → add citations → revise the prose
And then submit it as his or her own work. That’s not really a student learning to write a term paper. That’s a student learning how to operate a term-paper machine.
And that may be the more important story—not that AI is making college students dumber, but that we may be changing what it means to be educated if students routinely outsource the difficult parts of thinking.