Artificial intelligence has moved from novelty to infrastructure in education faster than almost anyone predicted. Across my peer-reviewed research — much of it co-authored with Subhashini Sumanasekara — a consistent picture emerges: AI's promise in education is real, but it is conditional on how thoughtfully we deploy it.
Personalized learning at scale
The most cited benefit of AI in education is personalization. Adaptive systems can adjust pace, difficulty, and examples to an individual learner in ways a single teacher managing thirty students cannot. Our research on AI-powered personalized learning found genuine promise here — but also pitfalls, including over-reliance on opaque algorithms and the risk of narrowing rather than expanding what students explore.
Reaching neurodiverse and special-education learners
Some of the most encouraging findings concern special education. AI tools can provide patient, repeatable, judgment-free support for neurodiverse students, adapting to communication styles and offering multiple representations of the same concept. The technology does not replace specialist educators; it extends their reach.
Emotional AI and student motivation
A growing body of work examines "emotional AI" — systems that detect and respond to a learner's engagement or frustration. Our review of emotional AI for motivation and retention found early promise for keeping students engaged, alongside serious questions about privacy, consent, and the ethics of machines reading children's emotions.
The infrastructure problem
A theme that recurs across this research is that AI's benefits depend on infrastructure — devices, connectivity, teacher training. In under-resourced contexts, AI can widen gaps rather than close them. Technology layered onto inequality tends to amplify it.
Ethics cannot be an afterthought
Student data, algorithmic bias, and informed consent are not edge cases; they are central. Our work on student-centric ethical frameworks argues for participatory consent and genuine data ownership, especially when the learners are children who cannot meaningfully consent for themselves.
Final thought
AI will not replace teachers, but it is already changing what teaching looks like. The evidence suggests the technology is powerful and double-edged: transformative where deployed with care and equity, harmful where deployed carelessly. You can read the underlying studies on my research page.