Intelligent Tutoring Systems
How AI-powered tutors simulate 1-on-1 instruction with effect sizes comparable to small-group teaching.
ITS Feature Comparison
Compare 6 intelligent tutoring systems across subjects, AI techniques, evidence levels, and key features. Sort by any column, expand rows for details, and filter by subject or evidence quality.
Research Evidence Cards
Interactive cards presenting key research findings on ITS effectiveness. Filter by evidence quality, click to reveal full details, and generate a research summary of the strongest evidence.
Key Insight
In 1984, Benjamin Bloom demonstrated that students receiving 1-on-1 tutoring performed 2 standard deviations above conventionally taught peers — meaning the average tutored student outperformed 98% of the control group. This became known as Bloom’s 2-Sigma Problem: how can we achieve tutoring-level outcomes at scale? Intelligent Tutoring Systems are the most promising technology for closing that gap. Meta-analyses show ITS achieving effect sizes of around 0.76 sigma — better than conventional instruction and approaching the effectiveness of small-group human tutoring. That still leaves roughly 1.24 sigma between ITS and the gold standard of expert 1-on-1 instruction. With the emergence of large language models powering the next generation of tutoring systems, that gap may finally be narrowing — but educators should remain both hopeful and critically engaged with the evidence as it develops.