Wharton Finds Structured AI Improves Learning Outcomes

The Wharton School published a Feb. 24 study testing how AI assistance timing affects long-term skill development in a three-month experiment with more than 200 chess learners. Participants given system-regulated, scheduled AI guidance improved about 64% versus roughly 30% for those with unrestricted on-demand help, and retained gains weeks later. The findings suggest AI tutors should use calibrated, bounded support to foster deeper learning.