The worries that artificial intelligence is degrading education have been building for months. Students are using it for everything, including cheating, and a recent report out of M.I.T. said it is triggering “cognitive surrender.”
Many university leaders are embracing A.I. with enthusiasm anyway.
As A.I. leaders warn about the dangers of their own technology, the dissonance between the optimism of university leaders and the fears of many students and professors may increase. Professors and students are often in a bind about what they should do: Use more A.I. to stay relevant, or dial back to preserve learning and thinking?
While some politicians and A.I. leaders are warning about possible doomsday scenarios for humanity as A.I. models advance, many campus leaders are focused on the existential threat the technology poses to universities.
The committee at the Massachusetts Institute of Technology studying A.I. use released a number of alarming findings, including that the technology has caused “major shifts in campus culture,” isolating students who are skipping office hours and ditching communal studying. Getting answers from a bot can create “the illusion of learning” for students who reach for A.I. “at the first hint of struggle,” the M.I.T. committee wrote in its August report.
At the same time, “the potential of these technologies to augment work across campus is immense,” the committee wrote.
The M.I.T. report crystallized some of the thinking in higher education about A.I., but top administrators have been embroiled in debates for several years. Administrators have often been the most eager to champion A.I. efforts, while many faculty and students remain more skeptical.
“I think you have to lean in, but you have to do so in a way that’s thoughtful and ethical and asks the right questions,” said Wendy Hensel, the president of the University of Hawaii, where she presented an A.I. strategy to regents in July. “The right answer to A.I. is not just ‘no.’ It can’t be. It has to be approached with curiosity and understanding of the implications.”
The hodgepodge of policies universities have created to respond to A.I. often reflect the confusion about how to proceed — with some clear dos and don’ts, but often a lot of wiggle room for faculty and students to figure it out on their own.
Experts Warn That A.I. Harms Learning
For a motivated student, an A.I. chatbot can function as a sort of tutor. Chatbots can break down the steps in a math problem, surface unexpected sources for research papers and offer feedback on writing.
Still, a growing body of research suggests that A.I. use can also harm learning.
A recent study of 26,000 Chinese secondary-school students found that A.I. adoption led to an 18 percent increase in homework scores, but a 20 percent decline in scores on secure exams. The implication was that rushing through homework with help from chatbots degraded the deeper learning needed to perform well when the technology was not available.
One 2025 study out of M.I.T. found that people who used chatbots as writing assistants struggled to remember what they had written only minutes after finishing their essay. “You basically didn’t engage any of the memory networks that would actually help you remember,” said Nataliya Kos’myna, a M.I.T. researcher who worked on the study.
Other research has shown that people who have greater trust in A.I. and rely on it more are more likely to have lower cognitive engagement.
But A.I. can be hard to avoid. Chatbots are embedded within the Google Classroom suite, which is nearly ubiquitous in schools, and within many online learning systems used by universities.
Some tools block outright cheating or plagiarism and allow professors to toggle specific A.I. capabilities on and off. But the most popular chatbots among students will instantly complete entire assignments, including research papers, on a student’s behalf.
Many Universities Are Embracing A.I. Anyway
Sian Leah Beilock, the president of Dartmouth, has argued that preparing graduates to use A.I. is an existential requirement of institutions of higher education.
“In a world that promises to be defined by A.I., universities that fail to produce graduates who can use it productively will consign themselves to irrelevance,” she wrote recently.
Harvard’s president, Alan Garber, said in public remarks this month that A.I.-backed research is on the cusp of a boom that may drive the pace of knowledge faster than at any other time in human history. Dr. Garber has said that researchers in many fields are now unlikely to be at the leading edge of their disciplines without A.I.
Ohio State is integrating A.I. into every major, with the expectation that every student will be fluent in their field of study by graduation. The university has directed every department and major — from business and engineering, to English and veterinary medicine — to train students on how to use A.I. in their respective fields, and when not to.
“What is OK to offload to A.I., and what is not?” Ohio State’s president, Ravi V. Bellamkonda, said in an interview in July. “The answer to that depends on the field and depends on your stage of learning.”
Elizabeth Kiss, the president of Union College, said universities have to educate students on both the thoughtful use of A.I. and on the things A.I. doesn’t do, such as “lead teams, push the boundaries of knowledge.”
Still, Some Schools Are Creating Boundaries
The University of Chicago has banned A.I. in certain required social sciences courses, while the university’s law school banned phones, tablets and laptops in first-year classrooms. The university is one of a handful that recently adopted new academic integrity rules.
The law school at the University of California, Berkeley, adopted a policy forbidding the use of A.I. “to conceptualize, outline, draft, revise, and edit” work submitted for credit.
Cornell University has created a disciplinary process meant to correct A.I. cheating while avoiding black marks on students’ records. Now, students accused of minor misuse will be allowed to accept responsibility and complete a workshop and some reflective work.
Steve Jackson, the vice provost for academic innovation, said the program would not be used for plainly egregious misconduct. “This is a chance to sanction something, catch it, not let it go unnoted or to let it pass,” he added.
Like many institutions, Harvard does not have a central set of A.I. rules. David Deming, dean of Harvard College, wrote in a note to students arriving for the fall semester that he favors more A.I. where “it can deepen learning,” but that the school was also looking to A.I.-proof some assignments.
That would mean more in-class tests, said Dean Deming, and fewer take-home assignments, where the temptation to cheat is powerful. He said he intends to produce a set of guidelines for A.I. use at the college by the end of the academic year.
For younger students, some school districts are proceeding cautiously. Earlier this month, the nation’s largest school district, in New York City, vastly limited student A.I. use in grades K-8, while the second-largest district, in Los Angeles, essentially banned student use of chatbots at every grade level.
Policing A.I. Use Is Hit or Miss
Students are often navigating this landscape without clear guidance from their colleges.
Most institutions leave it up to individual instructors — who may have little experience with A.I. — to set rules on how the technology is used in each class.
It may seem clear that submitting chatbot-generated text is cheating, while using a chatbot to revise one’s own writing is allowed. But the line between those two capabilities is thin, and much A.I. use lies in an ethical gray zone.
Many educators “don’t even know where to start,” said Annie Chechitelli, chief product officer at TurnItIn, a company that makes software that can detect plagiarism and track and guide A.I use. “What is OK? What’s not OK?”
Many instructors have already moved to on-paper assignments in the classroom, or locked-down computer browsers that limit A.I. misuse. Others report success in using A.I. detection software, which has recently improved in its ability to flag chatbot-written text while avoiding false positives.
Some see A.I. as a crucial partner for thinking and writing and put few limits on how students use the technology. Others subscribe to the notion of “A.I. literacy,” assigning students to use chatbots to produce an essay or poem, for example, and then critique and try to improve the output.
In Australia, the University of Sydney has adopted the “two-lane approach,” seen as a global model. Across disciplines, students must complete some assessments in person, under secure conditions meant to restrict A.I. But other assignments must use A.I., to help familiarize students with the technology that is reshaping the economy.
“It’s almost like sex ed,” said Anita Sarma, a professor of computer science at Oregon State who has studied the impact of A.I. on student cognition. “They will do it anyway,” she said. “How do we teach them responsible use?”
The post Universities Sound Warnings About A.I. Even as They Embrace It appeared first on New York Times.
