New AI Tutor Achieves 0.71–1.30 SD Effect Size in Dartmouth Course — Study Results Published

What Happened

A scientific study from the InTextbooks 2026 Workshop (Utrecht) documents that a new AI tutor implemented in a Dartmouth University course achieved an effect size of 0.71–1.30 standard deviations in student learning outcome improvement. The study was shared on Hacker News where it scored 113 points.\n\n## Context and Impact
For comparison: typical 1-on-1 tutoring effect (Bloom's 2 sigma problem) is ~2 SD, while tech interventions typically achieve 0.2–0.4 SD. The 0.71–1.30 SD result is a breakthrough for AI education, suggesting AI tutors are approaching the effectiveness of human tutoring.\n\n## Details

• Study presented at InTextbooks Workshop 2026, Utrecht
• PDF available via Hacker News with high score of 113 points
• Measured real course outcome impact (not just satisfaction)
• Potential implication: if reproducible, AI tutors could democratize access to quality education