Publication
arXiv
Stage
Preprint
What we read
Summary of the paper
Authors
Curtis Northcutt, Inaara Hasmani, Kevin Feng, Trevor Khangi, Andreas Plesner, Jonas Mueller
Universities and research institutions
Not yet supplied in verified metadata; the Brief does not guess.

What they did and found

Across thousands of sessions, artificial intelligence tutoring produced Graduate Record Examination score gains similar to expert human tutors, with some AI tools matching or beating humans in certain topics, while cutting cost per point gained.

Why it matters

This could let schools offer more affordable, personalized help, but results vary by subject and not every learner will see the same gains.

In the study, one AI tutor reached gains close to a human tutor in several Graduate Record Examination areas. Overall, AI tutoring showed gains similar to humans.

The finding depends on the tasks tested and hour tutoring window; lower cost stands out, but tutor choice and response time affected engagement and results.

Read the paper PDF ↗

Original sources · 1
  1. StudentBench: AI and human tutoring yield equivalent GRE learning gains ↗arXiv · 2026-09-23

Check the original paper for its authors, methods, version and access terms.

What remains uncertain

Immediate gains were measured; long-term retention and results outside Graduate Record Examination testing or other languages were not assessed in this study.