AI Tutoring With Repetition Boosts Student Math Learning

A recent study involving over 6,000 middle school students in Tennessee suggests that artificial intelligence can enhance learning, particularly when it encourages students to slow down and master concepts. Researchers investigated four different approaches to practicing fractions, randomly assigning students to either conventional computer-based instruction or the same software augmented with an AI tutor. Within these groups, a subset of students were required to correctly answer practice questions demonstrating the same skill three times consecutively if they initially made an error. This AI-enhanced "mastery learning" approach, which emphasizes sticking with a skill until a certain level of proficiency is demonstrated, resulted in students scoring approximately 3 percentage points higher on a retention test compared to those receiving standard computerized instruction. The study, conducted using software developed by the researchers that mirrored platforms like Khan Academy, involved a single 50-minute session during math class, followed by a 15-minute assessment one week later. While the observed advantage was modest, it provides early evidence of AI's potential positive impact on learning, especially in contrast to growing concerns that AI might hinder education by providing answers directly and bypassing the learning process. Philip Oreopoulos, an economist at the University of Toronto and lead author of the study titled "Making AI Tutoring Productive: Evidence from a Mastery-Based Math Practice Experiment," cautioned against overstating the findings, stating that while it's not yet a definitive "game changer," it offers "hints that it has some positive value against no AI at all." The research specifically focused on the efficacy of AI tutoring when coupled with a mastery-based practice regimen, where students must prove competence before advancing. This contrasts with typical AI applications that might simply accelerate task completion without ensuring deep understanding. The study's design allowed for a direct comparison between AI-assisted learning with and without the mandatory repetition component, as well as a baseline of non-AI computer instruction. The results indicate that the combination of AI guidance and deliberate practice is key to unlocking its educational benefits, suggesting that the AI's role is not just to provide information but to structure and reinforce the learning experience. The experiment's scale and the use of a large, diverse student population in a real-world classroom setting lend weight to its conclusions regarding the nuanced application of AI in education.