That's the take of a leading scholar who has written extensively about the evolving role of AI on campus and efforts to establish suitable guidelines regarding its use.
“The technology is changing quickly, and institutions are still trying to figure out what adequate AI governance should look like in practice,” says Dr. Chuhao Wu, Strategic Assessment Librarian at Clemson University.
Wu tells The EDU Ledger that having a written AI policy is “only the first step.”
“The next question is whether the policy is actually useful to the people who need it,” Wu says. “Is the information easy to find? Is it organized in a way that students, faculty, and staff can understand effectively? I think this information architecture issue is still underexamined in studies of higher education AI policies.”
The Data Strategy Gap
Surveys support Wu’s analysis. An EDUCAUSE QuickPoll of institutions of higher education found that only 1% of respondents indicated that they felt their institution had fully modernized its data management and strategy.
The same poll found that only 30% of respondents indicated that preparation for AI and analytics was a priority. A separate study found that at least 77% of respondents had “at least some AI-related strategy in place at their institution.”
“Taken together, these findings point to potential gaps in AI-related data preparation,” wrote Sean Burns, corporate researcher at EDUCAUSE.
Burns noted that the least-commonly reported priorities were governance and compliance (27%), automation (17%), and advancement and fundraising (11%). Other priorities included finance, human resources, and reputation management systems, according to Burns.
High Profile Missteps
Incidents of AI gone awry or being misused in higher education are not hard to find – and students and faculty alike are becoming ensnarled in investigations into whether they used the technology to take shortcuts on their work.
For instance, in April 2025, Springer Nature published an article titled “Mastering Machine Learning: From Basics to Advanced” that contained “numerous purportedly nonexistent or materially incorrect academic citations,” according to the AI Incident Database.
“While most universities had guidelines for faculty and students, less than 20% of the universities provided GenAI guidelines for researchers,” researchers observed in a 2025 investigation of higher education institutions’ guidelines and policies regarding the use of generative AI.
A cheating scandal rocked a big computer science class at Purdue University in April 2026 after a professor accused students of using AI to cheat and threatened to fail them if they didn’t confess or deny the allegations. Relatedly, a 2025 study on generative AI in higher education found that while 41% of universities offered detailed guidance for AI use in the classroom, most of the guidance dealt with writing, whereas references to code and STEM-related activities were sparse.
Death of the Honor System
In May 2026, Princeton University scrapped its 133-year honor system and started proctoring exams amid concerns that students were using AI so frequently that it was becoming a burden on other students who were duty bound to report it.
For Daniel S. Schiff, both the Purdue scandal and the Princeton proctor move hit close to home. Schiff is an assistant professor in the department of political science at Purdue and co-director of the Governance & Responsible AI Lab, or GRAIL. He is also an alum of Princeton.
“It’s a big change. This is something that was important to the ethos of Princeton,” Schiff says of his alma mater’s old honor system that relied on students to report cheating. “I think it’s telling you that that’s the last straw for them,” he says of the Princeton move to start proctoring exams.
Schiff says the Princeton decision could be a bellwether for higher education.
“At minimum, we’re seeing such a high rate of misconduct among students that it doesn’t feel tenable for many, many people in institutions to not have students in the classroom with a proctor,” Schiff says. “Take-home exams are not gonna be trusted nearly as much anymore. That is a great indication of the moment.”
Schiff says the biggest issues he sees with the use of AI in higher education are misuse, overreliance and cognitive deskilling or misconduct.
“There’s a lot of cool stuff that AI can do in education for teaching and learning,” Schiff says, “but what’s happening most often is people skipping their work, skipping their homework, not learning, and it’s really threatening a lot of pedagogical models.”
Administration and Labor
Potential misuse or abuse of AI transcends what takes place in the classroom with coursework, teaching and learning. The use of AI is also being used as an administrative tool or being contemplated in other areas of higher education, such as in admissions and counseling or even hiring.
“Do you want to use an AI tool for hiring, or is this going to be unfair and create other problems,” Schiff asks. “So there has to be an organizational posture toward responsible adoptions.”
In 2025 Illinois became one of the fi rst states to prohibit the use of AI in therapy and psychotherapy. A top Illinois offi cial says the law is meant to ensure that people get quality healthcare from “real, qualified professionals” instead of computer programs that “pull information from all corners of the internet to generate responses that harm patients.”
A Roadmap for Leaders
Wu, the Clemson AI researcher, says while AI “does not appear to be causing large-scale job replacement in higher education at this point,” institutions are creating or recruiting AI-related roles, such as directors of AI initiatives or AI literacy librarians.
“These positions are generally not presented as replacements for existing workers,” Wu tells The EDU Ledger. “That said, I do think there can be a labor issue here. Many people are shifting part of their work toward AI, but their previous responsibilities have not disappeared.”
Subsequently, Wu says he would emphasize the need for formal recognition of AI-related labor.
“New responsibilities should be reflected in workload expectations, evaluation criteria, and, where appropriate, compensation,” Wu says, noting EDUCAUSE makes a similar recommendation in a 2026 report on the impact of AI in higher education that says that institutions should codify new AI-related work responsibilities in job descriptions.
Wu says he would emphasize three points about AI for higher education administrators.
“First, I would encourage administrators to invest in people,” Wu says. “I know many institutions are already doing this by hiring new experts and providing training for employees. But I am also aware that many people are shifting their work focus toward AI, while their original responsibilities have not been reduced or reallocated. That may not be sustainable in the long term.”
Wu says he would also emphasize the need for campus-wide coordination.
“AI governance should not mean that every decision is centralized. Different units will still need flexibility because they serve different needs,” Wu says. “But institutions should have a clear way to share information, align priorities, reduce duplicated efforts, and make sure that important issues are not handled inconsistently across campus.”
Wu’s third and final point is to remind people that the core objective of higher education is “not simply to teach students how to use a new tool.”
“The goal is to cultivate critical thinking and foster lifelong learning,” Wu says. “AI is not merely a tool that students need to learn to use. It is also a major technological development that students, faculty, and staff will continue to question, shape, and be shaped by.
“Therefore, I believe the fundamentals of AI governance should not be only AI-oriented,” Wu says. “They should be oriented toward helping students and workers make sound decisions about technology, themselves, and society.”
Jamaal Abdul-Alim is a longtime education writer who resides in Washington, D.C. He occasionally teaches journalism and has completed fellowships at the University of Michigan, Columbia University and New America.
This story originally ran in the July issue of The EDU Ledger’s 2026 Summer Supplement.
