EdTech Research
AI-Enabled Adaptive Learning: A Comprehensive Review
November 2025 · 8 min read
Artificial intelligence is fundamentally transforming how educational content is delivered and personalised. A comprehensive 2025 review by Tan et al. examines the landscape of AI-enabled adaptive learning platforms, revealing both remarkable progress and important considerations.
The review analyses dozens of platforms and research studies, identifying key patterns in how AI is being used to personalise education at scale.
The evolution of adaptive learning
Adaptive learning has evolved from early rule-based systems to platforms capable of modelling individual learners in real time. The latest generation analyses patterns in student behaviour, predicts areas of difficulty, and adjusts content dynamically.
Unlike earlier systems built on pre-programmed decision trees, modern platforms can discover unexpected patterns in learning data, surfacing intervention opportunities a human instructor might miss.
Key capabilities
The review identifies the core capabilities that distinguish effective platforms: intelligent content sequencing that follows learner performance; predictive analytics that flag at-risk students early; meaningful feedback on student responses; and dashboards that give educators actionable insight.
Particularly promising is the integration of multiple techniques — combining knowledge tracing (what a student knows) with signals about engagement lets platforms adapt to motivation as well as cognition.
Effectiveness evidence
The literature consistently shows positive effects, though sizes vary with implementation quality. The largest benefits appear when AI adaptation is combined with quality content and ongoing teacher involvement.
AI-enabled platforms appear to particularly benefit students who might otherwise struggle in traditional settings — immediate, personalised support catches gaps before they compound.
Challenges and future directions
The review also names the hard parts: equity of access, student privacy, transparency in AI decision-making, and avoiding over-reliance on technology at the expense of human connection.
Future directions include richer modelling of learning processes and AI systems that can explain their recommendations in ways teachers and students actually understand.
Key takeaways
- AI-enabled adaptive learning is a step change over rule-based systems
- Most effective when combined with quality content and teacher involvement
- Particularly benefits students who struggle in traditional environments
- Privacy, transparency and equity are critical implementation considerations
- Future platforms will blend multiple AI techniques for holistic personalisation
What this means in practice
For schools comparing platforms, look past surface features to how AI is actually used: meaningful teacher oversight, transparency about decisions, and a focus on learning outcomes over engagement metrics. That is the standard Lumen8 holds itself to — including an assistant that only ever proposes, never acts alone.
Source: Tan LY, Hu S, Yeo DJ, Cheong KH. Artificial intelligence-enabled adaptive learning platforms: A review. Computers and Education: Artificial Intelligence, Vol 9, 2025, 100429. doi: 10.1016/j.caeai.2025.100429.