Is artificial intelligence sufficiently ready for the Turkish dentistry-specialty-exam?

Abstract

The main purpose of this research is to examine the impact of artificial intelligence (Al) on the field of education, especially the relationship between the success of candidates in the dentistry-specialty-exam and the success of artificial intelligence supported chatbots. We used ten questions prepared in a private educational institution called DUSDATA-Tr, which were used in the preparation process for the dentistry-specialty-exam. These ten questions were answered by 657 students and Al programs called ChatGPT, Bard, and Bing, which have Turkish language features. SPSS 25 (IBM, Corp.) was used for statistical analysis. Descriptive statistics for each variable were calculated and data were analyzed with Pearson chi-square test (p<0.05). The results showed that Questions 4 and 10 have the highest success rate, but Question 2 has the lowest. While the ChatGPT correctly answered the 1st, 5th and 10th Questions, the Barg correctly answered the 7th and 8th Questions, and the Bing correctly answered the 5th and 10th Questions. There was no difference between the correct and incorrect response rates of Al programs (p = 0.942). Question difficulty was negatively correlated with the Bing program (r =-0.34), (p<0.05). This research has revealed differences in the accuracy of different AI-programs' responses to educational materials and specific questions. Although the statistical results show that these differences are not statistically significant, they provide important insights into the development of Al-programs and their potential use in education.

Ähnliche Arbeiten