Primed but sceptical: Student engagement, performance, and perceptions of AI-generated lecture primers
This study investigates whether written, audio, human discussion, or AI-generated lecture primer formats differentially affect student engagement, academic performance, and perceptions of efficacy among first-year undergraduate finance
This study investigates whether written, audio, human discussion, or AI-generated lecture primer formats differentially affect student engagement, academic performance, and perceptions of efficacy among first-year undergraduate finance students. Engagement was universally low declining sharply after week one. Whilst no differences in engagement were observed in week one, Audio attracted significantly greater engagement than AI Discussion by week two, consistent with evidence that students remain sceptical of AI-generated content. Primer group allocation was not associated with differences in academic performance, supporting the integrity of group assignment. Sustained engagement beyond week one was associated with better performance: those engaging with AI Discussion and Audio primers after week one performed better on the mid-semester test and final examination respectively, suggesting these formats appealed disproportionately to higher-performing students. Perceptions of accuracy were highest for Audio, whilst AI Discussion users were least confident in content reliability. Importantly, for educators considering AI-generated preparatory materials all primer formats were rated equally and positively across key efficacy dimensions, suggesting students valued priming itself over delivery format. This suggests that AI-generated primers represent a viable and scalable alternative to human-produced equivalents, though transparency about their provenance remains advisable given residual student scepticism about reliability.
