Instructor-designed AI improves learning in a randomized trial across 19 universities

Impact 6 arXiv cs.AI

Summary

  • A randomized controlled trial across 7,200 students at 19 universities evaluated AI tools designed by frontline instructors.
  • Instructor-designed AI innovations improved student learning as measured by proctored, in-person examinations.
  • The trial found no significant change in student engagement or satisfaction metrics compared to standard practices.
  • The findings suggest AI can enhance educational outcomes when adoption is driven and customized by instructors for their classrooms.

Why it matters

This research provides empirical evidence that instructor-customized AI solutions can successfully improve educational outcomes, contrasting with the negative impacts often seen with generic or centrally mandated chatbots.

Who should care