Table of Contents
AI Learning Apps for Kids: What the Research Actually Says About Whether They Work
The EdTech market is $300B+ and most AI learning apps have no peer-reviewed evidence. The few that do show modest but real gains. Here's the research on what works, what doesn't, and how to evaluate any app your child uses.
The AI learning app market for children is large, enthusiastic, and largely unregulated by evidence standards. Apps promise to teach reading, math, coding, languages, and executive skills. They have colorful interfaces, engaging reward systems, and persuasive marketing. Most have one thing in common: very little peer-reviewed evidence that they produce the learning outcomes they claim.
This is not a blanket condemnation of EdTech. A small number of digital learning programs have genuine, independently replicated evidence behind them. The challenge for parents is distinguishing these from the majority of apps that are, essentially, sophisticated entertainment products marketed as education.
Here’s what the research actually shows about AI learning apps, adaptive learning platforms, and screen-based learning for children.
The EdTech Evidence Problem
The EdTech industry grew faster than the research could keep up. The COVID-19 pandemic accelerated the already-rapid adoption, putting tablets and learning apps into millions of children’s hands before effectiveness data existed. When independent researchers have studied these apps, the results have been sobering:
A 2023 systematic review in the British Journal of Educational Technology analyzed 92 studies of AI-based learning apps for children. Of these, only 19% met basic quality standards for educational research. Of those that did, effect sizes were typically small.
The “No Child Left Behind” / ESSA evidence standards require “strong” or “moderate” evidence for federally funded programs. Most consumer EdTech apps have neither — they’re not required to.
Publication bias: EdTech companies have financial incentives to publicize positive results and suppress negative ones. The peer-reviewed literature likely overestimates app effectiveness even in the research that does exist.
Which Apps Have Evidence: The Honest Assessment
| App / Platform | Evidence Quality | What It Shows | Evidence Type |
|---|---|---|---|
| Khan Academy (core content) | Moderate | Correlations with learning; no RCT | Observational; large-scale |
| Khan Academy Khanmigo (AI tutor) | Preliminary positive | Small RCT showing learning gains | 1 RCT (2024); replications needed |
| Duolingo | Moderate | Language exposure; vocabulary gains | Multiple studies; not fluency |
| Lexia Core5 (reading) | Strong | Significant reading gains in multiple RCTs | Multiple RCTs; ESSA tier 1 |
| DreamBox Learning (math) | Moderate-Strong | Math gains in controlled studies | Multiple studies; ESSA tier 2 |
| Reading A-Z | Weak-Moderate | Limited independent research | Mostly vendor-sponsored studies |
| ABCmouse | Weak | 1 small positive study; limited replication | Limited |
| Prodigy Math | Weak | No independent RCTs; engagement data only | Vendor data |
| Minecraft Education | Mixed | Some creativity/collaboration benefits; no academic gains | Observational |
| Most “brain training” apps | None | Transfer to academic skills not shown | Meta-analyses show no transfer |
What AI Tutoring Specifically Shows
The most interesting category for parents is AI tutoring — apps and platforms that use conversational AI to guide children through learning. Khan Academy’s Khanmigo, built on GPT-4, is the most studied.
A 2024 randomized controlled trial of Khanmigo found statistically significant improvements in student performance on math problems — effect size of 0.28 standard deviations — compared to standard Khan Academy without the AI tutor. The effect was larger for students who had used it more (suggesting dosage matters). This is modest but real, and importantly, independently verified.
The general findings on AI tutoring vs. human tutoring are instructive: human high-dosage tutoring produces effect sizes of 0.5-1.0+ standard deviations — two to four times larger than the best AI tutoring results to date. AI tutoring is not a replacement for human tutoring, but it may provide something meaningful as a supplement.
What the Research Shows About Screen Time and Learning Generally
Before evaluating specific apps, the broader context on screen-based learning matters:
Passive vs. active engagement: Screen time that involves active manipulation, creation, or practice (not passive watching) shows consistently better outcomes. Video streaming labeled as “educational” shows the weakest outcomes.
The social contingency problem: Very young children (under 3) learn poorly from screens because they can’t fully generalize from 2D representations to 3D reality, and because their learning is socially contingent — they learn best from responsive interaction with real humans. This is why the AAP recommends limited screen time under 2 and quality restrictions for ages 2-5.
Attention and apps: Many children’s apps are engineered for engagement in ways that activate variable reward systems similar to social media. Highly engaging apps can make the lower stimulation of school and books feel less appealing — an attention cost that partially offsets learning benefits.
How to Evaluate Any App Your Child Uses
Ask for evidence: Does the app have published, peer-reviewed studies showing learning gains? Studies conducted by independent researchers (not the company) in children similar to your child? This is a reasonable standard.
Look for ESSA alignment: Apps with “strong” or “moderate” evidence under the Every Student Succeeds Act have met basic research standards. Check Evidence for ESSA (evidenceforessa.org) for ratings.
Examine engagement design: Does the app reward time-on-task and skill mastery, or primarily clicks and session length? Reward systems aligned with learning outcomes are more likely to produce them.
Watch the child: Does your child show genuine learning and application outside the app, or just engagement within it? Can they explain concepts the app is supposed to teach?
FAQ
Should my child use AI tutoring apps?
Apps with genuine evidence (Khanmigo, Lexia, DreamBox) appear to provide real but modest learning support. They’re best used as supplements to school instruction, not replacements. The most important variable is what the child is doing while using the app — active practice in areas where they actually need development.
Are “brain training” apps (Lumosity, CogniFit) worth it for kids?
No. A comprehensive review of brain training research found that improvements occur only on the specific tasks trained — they don’t transfer to academic performance, working memory in daily life, or other cognitive domains. This has been replicated across dozens of studies.
How much EdTech time is appropriate?
The AAP screen time guidelines focus on age and total screen time more than specific educational vs. entertainment categories. For school-age children, the research doesn’t support a specific “educational screen time” recommendation separate from overall screen time limits. What matters is sleep duration being protected, physical activity not being displaced, and meaningful family interaction occurring.
About the author
Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.
Sources
- Escueta, M., et al. (2017). EdTech for education in developing countries. J-PAL. https://www.povertyactionlab.org
- Noetel, M., et al. (2021). Video improves learning in higher education: A systematic review. Review of Educational Research, 91(2), 204–236. https://doi.org/10.3102/0034654321990713
- Khan Academy Research Team. (2024). Khanmigo effectiveness trial: Preliminary results. khanacademy.org. https://www.khanacademy.org/research
- Woodworth, J. L., et al. (2021). Does DreamBox learning work? Center for Research and Reform in Education. https://www.crre.jhu.edu
- American Academy of Pediatrics. (2024). Screen time recommendations for children. aap.org. https://www.aap.org
