What happened
EdSurge columnist Mike Kentz says optional AI tutors may not prompt student effort. A National Bureau of Economic Research working paper by economists Philip Oreopoulos and Nina Low tracked Khanmigo across 18 Tennessee middle schools. Although 96% tried it, median use was one-third of available days, and math gains were modest versus usual remedial instruction.
Why it matters
Access alone is a weak proxy for learning: students may use a tutor infrequently or not for the target subject. Kentz argues that requiring AI can address participation, not effort; assessing students’ conversations may make their reasoning more visible. His separate 21-student pilot reported shifts in students’ approaches, but had no control group and relied mostly on self-reports. It is preliminary evidence, not proof of learning gains.
What to do next
If you want students to use AI for disciplinary thinking, decide first whether it belongs in the course; Kentz notes that teaching without AI remains a legitimate choice. For a trial, set a bounded task requiring students to explain, question or respond in a role-play, then assess transcripts with a rubric for subject understanding and reasoning. Prepare students and review interactions, adjusting difficulty: classroom beta feedback included reports of repetitive or interrogating exchanges. Treat early results cautiously, distinguishing participation from learning before expanding use.
About this briefing
Reviewed by TutorFlow Editorial. We link the primary source, preserve its publication date, and distinguish reported claims from TutorFlow analysis. Our commentary focuses on practical decisions for educators and training teams.
Discussion (0)
Open to everyone