AI versus human-designed exam questions in undergraduate dental education: A randomized controlled study

TL;DR

A study compared AI-generated and human-designed exam questions in dental education and found no performance difference, but students preferred human questions for clarity.

Contribution

This is the first randomized controlled trial comparing AI-generated and human-designed exam questions in undergraduate dental education.

Findings

No significant difference in student performance between AI-generated and human-designed questions.

Students preferred human-designed questions for clarity and familiarity.

Abstract

The comparison of the effectiveness and fairness between AI-generated and human-designed exam questions in the context of undergraduate
dental education is relevant. Therefore, it is of interest to examine the use of AI-generated versus human-designed exam questions in
undergraduate dental education. A randomized controlled trial involving 289 students was conducted to compare the effectiveness and
fairness of both formats. Results showed no significant difference in overall student performance between the two groups. However,
students preferred human-designed questions for their clarity and familiarity. Thus, the role of AI-generated exam questions in
undergraduate dental education compared to traditional human-designed questions remains a challenge.