The ads promise a personalized teacher for every child — an AI that adapts to their level, fills their gaps, and never loses patience. The skeptics say it is screen time dressed up as education. You have probably tried at least one adaptive learning platform and wondered the same thing every parent wonders: is this actually working, or is my kid just clicking through?
The research has a surprisingly clear answer. AI tutoring systems genuinely work — the effect sizes are large and the evidence is robust across decades of studies. But the research also reveals a paradox that no one in the marketing department will tell you: students learn more from AI tutors while feeling less satisfied with the experience. Use AI for what it does best and keep what only you can do.
The Evidence Is Stronger Than You Think
Kurt VanLehn (2011, Educational Psychologist) compared human tutoring, intelligent tutoring systems (ITS), and other instruction. ITS effect size ~d = 0.76 vs human tutoring d = 0.79 (VanLehn, 2011). The gap was negligible.
Kulik and Fletcher (2016) meta-analysis of 50 evaluations: median ITS effect d = 0.66 (Kulik & Fletcher, 2016). Effects were larger on locally developed tests than standardized tests.
The Paradox: Learns More, Feels Worse
Kestin et al. (2025) found students using an AI tutor learned more in less time than in-class active learning (Kestin et al., 2025), but felt less confident and less satisfied. Discomfort does not mean the tool is failing.
What Is Available Now
Tan et al. (2025) on adaptive platforms combining practice with immediate feedback (Tan et al., 2025). Wang et al. (2023): adaptive systems improved outcomes but teachers still mattered for motivation (Wang et al., 2023).
How to Use This at Home
Adaptive learning adjusts difficulty based on performance. ITS gives personalized instruction, practice, and feedback. Framework: Supplement, Don’t Replace
AI Does Best | You Do Best |
|---|---|
Adaptive practice (adjusts to level) | Introducing new concepts with context |
Immediate feedback on errors | Motivation and emotional support |
Gap-filling and remediation | Making learning meaningful |
Spaced repetition for retention | Relationship and accountability |
Tireless patience with drill | Knowing when to push and when to stop |
Tools: Khan Academy, IXL, DreamBox Math, Khanmigo.
What to Avoid
The Outsource Everything Trap: handing a child a tablet and walking away. AI is practice, not a replacement for you.
Resources
Khan Academy, IXL, DreamBox Math (K–8), Khanmigo.
Weekly question: Are you using any AI learning tools in your homeschool? Reply and tell me which ones — and whether you think they are actually working, or just keeping your child busy.
References
[1] VanLehn 2011 — https://www.tandfonline.com/doi/abs/10.1080/00461520.2011.611369
[2] Kulik & Fletcher 2016 — https://journals.sagepub.com/doi/abs/10.3102/0034654315581420
[3] Kestin et al. 2025 — https://www.nature.com/articles/s41598-025-97652-6
[4] Tan et al. 2025 — https://www.sciencedirect.com/science/article/pii/S2666920X25000694
[5] Wang et al. 2023 — https://www.tandfonline.com/doi/abs/10.1080/10494820.2020.1808794