Private school network bets big on AI tutors despite research gaps

Alpha School, a private school network, is expanding to about 50 campuses this fall, relying on AI tutors for two hours of daily instruction. The company trains its models on student misconceptions and learning differences, but researchers caution that the effectiveness of this approach is not yet scientifically established. Tuition ranges from $40,000 to $75,000 per year.

Alpha School's daily structure dedicates morning hours to AI-driven instruction, with students extending that time if they fail to hit performance benchmarks. Afternoons shift to workshops in coding, entrepreneurship, and public speaking. The company's learning scientists collaborate with software developers to refine models using current research and curriculum tools, with tuition ranging from $40,000 to $75,000 annually.

The expansion follows a 2025 Harvard randomized study in introductory physics where adaptive AI tutoring produced median learning gains more than double those of active learning methods. However, Alpha has not released its own underlying data. Education researcher Dan Goldhaber noted the system resembles established platforms like IXL Learning, which have been publicly studied. The company's billionaire principal, Joe Liemandt, has stated ambitions to reach a billion students.