Students Learn More In Less Time With AI Tutor, Harvard Study Finds

Last Updated:September 29, 2026, 1:03 pm IST

Students using the AI tutor recorded a median post-test score of 4.5, compared with 3.5 among those who learned through the in-class active learning lessons, Harvard study found.

The study involved 194 UG students enrolled in Harvard’s introductory physics course. (AI Image)

Students learned more in less time when taught through a specially designed artificial intelligence (AI) tutor than through an in-class active learning session, according to a randomised controlled trial conducted at Harvard University and published in Scientific Reports.

The study involved 194 undergraduate students enrolled in Harvard’s introductory physics course. Researchers compared learning outcomes when students studied the same material through a custom AI tutor and through conventional in-class active learning. Each student experienced both methods during the two-week study.

Students using the AI tutor recorded a median post-test score of 4.5, compared with 3.5 among those who learned through the in-class active learning lessons. The researchers also found that the median learning gains in the AI group were more than twice those recorded in the classroom group. The difference was statistically significant.

The AI group also spent less time on the lessons. While students in the classroom condition had about 60 minutes for learning after accounting for testing time, the median time spent by students using the AI tutor was 49 minutes. About 70 per cent of students in the AI group spent less than 60 minutes on the material.

Students also reported higher engagement and motivation with the AI tutor. On a five-point scale, the AI group reported an average engagement score of 4.1, compared with 3.6 for the in-class group. Their average motivation score was 3.4 compared with 3.1 among students in the classroom condition.

According to the study, AI tutors could potentially be used to introduce difficult material before class, allowing classroom time to focus more on higher-order skills such as advanced problem-solving, projects and group work. They also suggested that AI tutoring could support homework, exam preparation and remedial learning.

The researchers noted that the results do not establish that structured AI tutoring will outperform in-class active learning in every setting, particularly when learning requires complex synthesis of multiple concepts or higher-order critical thinking.

Quick Answers

Powered by

Ask News18

How did students using an AI tutor perform compared to classroom learning in the Harvard study?

Students using the AI tutor achieved a median post-test score of 4.5 and experienced learning gains more than twice those of the classroom group, whose median score was 3.5. Additionally, the AI group spent less time on the material, averaging 49 minutes compared to about 60 minutes for the classroom condition.

What levels of engagement and motivation did students report while using the AI tutor?

On a five-point scale, students in the AI group reported an average engagement score of 4.1, compared to 3.6 for the in-class group. Their average motivation score was 3.4, compared with 3.1 among students in the classroom condition.

What potential educational applications for AI tutors were suggested by the Harvard researchers?

Researchers suggested that AI tutors could introduce difficult material before class so classroom time can focus on higher-order skills like advanced problem-solving, projects, and group work. They could also support homework, exam preparation, and remedial learning, though they may not outperform active learning when tasks require complex synthesis or critical thinking.

Powered by

Ask News18

About the Author

Education.Careers Desk

A team of reporters, writers and editors brings you news, analyses and information on college and school admissions, board and competitive exams, career options, topper interviews, job notifications, …Read More

Read More