The First American College to Make AI Fluency a Graduation Requirement

When you drive through West Lafayette on a winter morning, Purdue University makes its presence known in the same way that big Midwestern research universities do: with expansive brick buildings, a campus that resembles a small city, and engineering students moving purposefully between classes. It’s not a place that moves fast just to look current. Practical, long-lasting knowledge is the foundation of its identity.

Engineering in agriculture. aviation. production. The kind of knowledge that makes things grow and be built. Therefore, it carried a certain amount of weight when Purdue’s Board of Trustees approved what the university refers to as an AI “working competency” graduation requirement in December 2025. This was the first such mandate at any American university. This was not a tech school on the coast following a fad. Purdue was discussing the current definition of competence.

All undergraduates on the West Lafayette and Indianapolis campuses will be subject to the requirement, which will go into effect for incoming students in the fall of 2026. By the time they cross a stage to pick up their diploma, Purdue graduates will have to show that they can use current AI tools in their field effectively, comprehend the limitations of those tools, acknowledge the impact of AI on decision-making, and defend decisions that were influenced by AI-generated insights. The mandate is integrated into required coursework across departments; 35 courses have already been identified to carry the requirement forward. It is not delivered through a single stand-alone course or a separate exam. The strategy is intentional. The goal behind it has been made clear by Purdue’s senior vice provost for academic and student success, Haley Oliver-Jischke. “You will lose out on opportunities if you don’t understand it or know how to utilize it and apply it effectively,” she stated. It’s important to note that she is not a computer science administrator discussing AI; rather, she is a microbiologist by training. It’s a person who has seen firsthand how technology is changing her own field of study.

When you consider the student data that has been gathered, the timing is not shocking. The plagiarism detection platform conducted a survey in 2025. Approximately 90% of college students reported using AI for academic purposes, with over half utilizing it daily or multiple times a week, according to Copyleaks. It is used by about half to create outlines. 44% to produce real drafts. One in three will summarize what they have read. These figures represent the typical behavior of the modern undergraduate; they are not anomalous behaviors. Purdue is addressing the disconnect between that degree of informal use and a true comprehension of the tools’ true functions, their shortcomings, and how to evaluate their output critically.

One could legitimately argue that Purdue is merely catching up to what students are already doing independently and turning informal behavior into formal coursework. There is some validity to that argument. However, the way the university has framed the requirement—not as digital literacy in the traditional sense of knowing how to use software, but as a professional and ethical competency—is also something more intriguing. In their work, graduates must be able to identify “the presence, influence and consequences” of artificial intelligence. The wording is more expansive than it seems. It implies that comprehending AI goes beyond simply being able to successfully prompt a chatbot. It’s about recognizing when one is influencing you.

Purdue was the first to move in this direction, but it’s not the only one. Midway through 2025, Ohio State University announced its AI Fluency initiative, which calls for graduates starting in the class of 2029 to show proficiency with the technology. Both are public land-grant universities that were established in the same time period and are based on the same idea of practical education fulfilling societal needs. They are also coming to similar conclusions about what practical education entails in 2026. It’s important to notice the parallel. When land-grant universities begin to take action, it usually indicates a more general change in the way American higher education views preparation for the workforce.

The criticisms are valid and should be taken seriously. The requirement has been referred to by some observers as “another box to check,” raising concerns that it will be applied superficially. Others have questioned whether universities requiring AI proficiency are unintentionally ignoring concerns about how AI affects intellectual independence, critical thinking, and the creative process. These objections are not paranoid. They are real, and the truth is that it’s still unclear if Purdue’s method will result in graduates who truly comprehend the technology or just graduates who know how to use it with confidence without actually understanding what it’s doing. When students begin to arrive on campus in the fall, what the embedded curriculum actually teaches will serve as evidence that this distinction is important.

As this develops, there’s a sense that the debate over calculators in the classroom from the 1980s will gain a lot of traction in the coming ten years. The answer to the question of whether allowing students to use calculators would result in pupils who were unable to think numerically was complex and context-dependent back then. AI might take a similar route. Whether or not students should use it is not the question. They already do. The question is whether they can learn how to use it effectively through higher education—critically, skeptically, and with a true grasp of its capabilities and limitations. Purdue has made the decision to try. Others will come after. The policy is not the true test. It’s the classroom.