IDLab Seminar: Do AI Tutors Help Students Learn Better?
On February 6, 2026, the International Laboratory of Intangible-driven Economy (IDLab) hosted a research seminar focusing on the practical application of artificial intelligence in education.
Egor Ivanov, Junior Research Fellow at the laboratory, presented his paper, "The Effectiveness of AI Tutors: Preliminary Empirical Findings," based on an experiment he conducted.
As part of the study, the author designed and implemented an AI tutor into the Microeconomics course for first-year undergraduate students of the International Bachelor's Programme in Business and Economics at HSE University - Perm.
"We designed the tutor based on principles of evidence-based learning. It regularly prompts students to retrieve material on their own, asks clarifying and follow-up questions, revisits topics through spaced repetition, and gradually increases complexity to deepen understanding without causing cognitive overload. As a result, it serves as a 24/7 personal assistant that doesn't just explain concepts, but helps students genuinely retain and apply knowledge. The next step is to link the tutor's responses directly to course assignments and integrate graph-based exercises, as visualization is critical for understanding economics."
During the presentation, the speaker focused on assessing how the introduction of a digital mentor affects students' actual academic performance and engagement. To address this research question, a unique dataset was assembled, combining aggregated statistics on student interactions with the AI and their academic outcomes. Additionally, individual student characteristics—such as study habits, skills, and motivation levels—were taken into account.
The study developed a model to estimate the contribution of the AI tutor to students' academic success, focusing on two key aspects:
-
Direct relationship: how the intensity and patterns of interaction with the system correlate with final grades in microeconomics.
-
Consideration of individual differences: the extent to which student personality traits, cognitive skills, study habits, and motivation levels enhance or mitigate the effect of using AI.
The findings from this study aim to shed light on the conditions under which AI tutoring is genuinely effective. Moving forward, these insights will help developers of educational technologies create personalized solutions tailored to individual student needs.
