Alpha School, an AI-driven private network charging up to $75,000 annually and backed by figures including Reid Hoffman, Bill Ackman, and former Education Secretary Linda McMahon, reports that only 42% of its students achieved its core promise of learning twice as fast in both math and reading, according to internal data reviewed by The New York Times. The model replaces traditional teachers with "guides" who provide motivation while students use the TimeBack app for two hours of daily academics. Math outcomes are the primary shortfall: roughly half of K-8 students missed the 2x growth target, and only 37% of students initially at or below the 50th national percentile in math met the goal by winter. Guides and parents told Wired that students often worked late into the night to meet targets, causing stress, while former employees informed 404 Media that the AI generated faulty lesson plans, sometimes doing "more harm than good," and scraped external courses without permission. Alpha is expanding into public schools, with over 2,600 students in Texas and South Carolina now using the platform, including a pilot in Houston ISD where Superintendent Mike Miles noted students spent more time on the apps than planned. External research complicates the AI-tutoring thesis: a study of 26,811 Chinese secondary students found homework scores rose 18% with generative AI use, but closed-book exam scores fell 20%, and OECD PISA data showed daily AI users scored lower in science than non-users.
Read original
hackernews