A two-year cluster randomized trial in 18 Tennessee middle schools tested Khanmigo, an AI tutor from Khan Academy, during daily remedial math sessions. Students using Khanmigo showed math achievement gains of 1.3 national percentile ranks per term, or 0.06 to 0.08 standard deviations annually. Full-year engagement yielded up to 0.14 standard deviations improvement.
These gains closely matched Khan Academy practice without AI assistance. Analysis revealed low engagement: while 96% of students tried Khanmigo, the median student messaged it on only one-third of practice days and in just 17% of mistake-making exercise sessions. Most interactions were minimal—students sent bare answers or clicked suggested prompts rather than engaging in substantive dialogue.
The study concludes that realizing AI tutoring's potential depends less on access and more on boosting student engagement. Keywords include Artificial intelligence, AI tutoring, Khanmigo, computer-assisted learning, and educational technology.
来源: Hacker News · 由HeadlinesBriefing整理摘要