Analytics

• LMS trace data, early warning systems and their realistic accuracy
• False positives and the harm of labelling a student at risk wrongly
• Prediction that leads to support, not to surveillance

Personalisation

• Adaptive learning systems, mastery pathways and the evidence base
• AI tutors and feedback generation, with the accuracy caveat
• Learning styles as a persistent myth that adaptive marketing still repeats

Design

• Retrieval practice, spacing and generative activity as the mechanisms that work
• Where AI genuinely lowers the cost of good pedagogy
• Assessment redesign when generated answers are freely available

Ethics

• Student data privacy, FERPA and equivalent regimes, and consent
• Algorithmic proctoring harms and the documented accessibility failures
• Transparency with students and evaluating a vendor claim with evidence