A preliminary exploratory study on AI chatbot–assisted learning in enhancing dental students comprehension of intraoral radiographic techniques
- Suresh Kandagal Veerabhadrappa, Jayanth Kumar Vadivel, Anand Marya, Seema Yadav Roodmal, Shaul Hameed Kolarkodi, Abedelmalek Kalefh Tabnjh, Vipin Kailasmal Jain, Thantrira Porntaveetus, Apitchaya Pruksametanan, Siddharthan Selvaraj
- http://10.3389/feduc.2026.1853688
Abstract
Background
Intraoral radiographic techniques (IRTs) are mandatory components of undergraduate dental training; however, their inherent complexity and reliance on conventional teaching may result in inadequate comprehension. Artificial intelligence–driven chatbots provide interactive, personalized learning and may enhance understanding when used alongside traditional instruction. This exploratory study evaluated the effectiveness of ChatGPT–assisted learning as a supplementary tool to improve dental students’ theoretical comprehension of IRTs.
Methodology
This comparative study included third-year BDS students who were randomly allocated into two groups of 24 each: ChatGPT–assisted learning (ChatGPT-AL) and self-directed learning (SDL), following a standardized lecture on IRTs. The ChatGPT-AL group studied IRTs interactively using standardized prompts, whereas the SDL group studied independently. Baseline and post-intervention theoretical knowledge were assessed in both groups. Intragroup and intergroup comparisons were performed using paired t-tests and one-way ANOVA, respectively, with p < 0.05 considered statistically significant.
Results
Intragroup comparison of the ChatGPT-AL group showed statistically significant improvements across several IRT components, particularly for the bisecting angle technique (BAT) and occlusal radiography, including principles and concepts, film placement, central x-ray beam positioning and angulations (p < 0.05). In contrast, the SDL group demonstrated reduced post-instructional scores for most components of the BAT, followed by the paralleling technique and bitewing radiography. Intergroup analysis revealed consistently higher post-intervention mean scores in the ChatGPT-AL group, with a statistically significant difference for the film holder component of the BAT (p = 0.044).
Conclusion
ChatGPT-AL demonstrated trends toward improved theoretical comprehension of IRTs compared with SDL and may serve as a useful supplementary adjunct to conventional teaching methods.