Ethan Carter
The increasing adoption of generative AI among students is contributing to a rise in grade inflation at universities, as highlighted in a recent working paper from the University of California, Berkeley.
Students utilize generative AI in three main ways: augmentation, where the technology assists with tasks such as research while the student does most of the work; reinstatement, which involves introducing new AI-driven assignments; and displacement, where AI fully automates tasks typically completed by the student, like essay writing. Each of these methods has the potential to boost grades, but only augmentation and reinstatement are linked to actual learning and skill development.
Certain academic assignments, particularly unsupervised take-home tasks, essays, and homework, present ideal opportunities for AI displacement. In contrast, proctored exams, oral presentations, and in-class discussions do not lend themselves to this automation.
In this study, UC Berkeley senior researcher Igor Chirikov examined more than 500,000 student-course enrollments across 84 departments at a major Texas university from 2018 to 2025. He discovered that increases in grades were predominantly found in courses characterized by a significant number of writing and coding assignments, where take-home tasks held substantial weight. This finding indicates that students are turning to AI to cheat on certain assignments to achieve higher grades. Overall, the research team observed a 30 percent rise in “A” grades in “AI-exposed courses” since the arrival of ChatGPT.
This trend is not surprising; it exemplifies a well-established application of generative AI. a student’s grade point average (GPA) plays a crucial role in their future, influencing admission into graduate programs and access to lucrative job opportunities. In a landscape where industries increasingly incorporate AI, often impacting the job market for new graduates, it is understandable that students are looking for ways to secure their futures.
What is noteworthy is that, four years after generative AI became a fixture in our lives, this study reveals that American universities have yet to adapt to its implications.
With the rise of AI-induced grade inflation, employers face greater challenges in identifying strong candidates among recent graduates, according to the research. More critically, this growing dependence on AI in education threatens to produce a workforce lacking essential skills.
“If AI replaces skill-building activities during education, students may graduate with diminished abilities in areas where AI excels, perpetuating a cycle that links AI in education with increased automation in the workplace,” Chirikov warns.
An academic framework that accommodates AI-driven grade inflation risks creating a workforce unprepared to perform fundamental job responsibilities, which could lead to heightened reliance on AI and further automation, potentially triggering the feared job crisis that some experts argue is already unfolding in various sectors.
Some universities are beginning to address this grade inflation issue, but the effectiveness of their proposed solutions remains uncertain. At Princeton, where approximately 30% of seniors confessed to cheating primarily through generative AI in a recent survey, faculty recently voted to repeal a 133-year-old honor code that permitted students to take in-person exams without faculty supervision.
Meanwhile, Harvard faculty are deliberating a proposal to limit A grades to no more than 20% of the class.
Daniel Mercer
