University of Maryland Randomized Trial Finds Students Offered a GPT-4o Course Tutor Scored About Four Points Lower

A University of Maryland randomized trial covered 2,379 undergraduates and 30 instructors in fall 2025. In matched sections of the same course, students offered a course-integrated AI tutor finished about four percentage points lower in final grades, or 0.37 standard deviations, while learning management system participation fell 0.90 standard deviations and page views and active days also declined. About 15% of students offered the tutor used it, and nearly 74% of requests sought information, explanations or answers.

This randomized trial provides evidence that a course-integrated AI tutor coincided with lower grades and lower platform participation, directly relevant to whether and how universities embed generative AI in teaching platforms. The researchers state the findings do not establish that purpose-built AI tutoring tools are not beneficial, and they do not establish that lower platform use caused the lower grades. The study used course grades rather than an independently administered assessment, students already had access to ChatGPT and Gemini, most instructors did not use the more guided tutoring mode, and the student survey response rate was low. Estimated grade losses were larger among first-generation students, but the researchers caution against overinterpreting subgroup results. Universities embedding AI tutors in courses need to assess which learning activities the tool replaces, such as office hours, course-material reading and asking instructors questions. Product and teaching teams should examine actual student usage: requests for answers dominate while requests for feedback are rare, which may weaken the tutor's guided value.