Inteligência artificial no desempenho de estudantes de odontologia: revisão sistemática e metanálise

Abstract

Artificial Intelligence (AI) has proven to be a growing and innovative tool in several areas of Dentistry, with promising potential. In the educational context, its use has also been discussed, but data in the literature have been controversial in relation to its real effectiveness in the performance of different tasks by dental students. Therefore, the present study aimed to perform a systematic review with meta-analysis, considering the following PICO question: What is the effectiveness of the use of artificial intelligence in the performance of tasks by dental students? This review followed the guidelines proposed by the "Preferred Reporting Items for Systematic Reviews and Meta-Analyses" (PRISMA) and was registered with PROSPERO. The search was carried out in the PubMed/MEDLINE, Scopus, Embase, Web of Science, Scielo, Cochrane Library, and gray literature databases. The following keywords were used: "artificial intelligence", "machine learning", "deep learning", "dental education", "dental training", "dental curriculum", "dental students" and "dental schools", combined through Boolean operators. The searches were imported into the "Rayyan QCRI" software and analyzed by two reviewers independently. The inconsistencies were analyzed by a third reviewer. Randomized experimental studies were included, which compared the performance of dental students with and without the use of AI tools. The meta-analysis was performed using the 'RevMan web' program, and the outcomes evaluated were sensitivity and specificity (diagnostic task), and scores in evaluations. Risk of bias was assessed using the Cochrane Risk of Bias tool version 2.0 (RoB 2.0)16 . The searches resulted in 495 records. After the removal of duplicates, 227 went through the initial screening process (reading of titles and abstracts), and 16 articles were selected for full reading. In the end, seven articles were included, which had their data extracted for analysis. The studies were published between 2020 and 2024, three from the United States and four from Europe. Six studies were related to tasks in Dental Radiology, and one to Digital Dentistry. Regarding the outcome’s sensitivity and specificity, the meta-analysis, despite favoring the groups that used AI, showed no significant effect (p = 0.13 / p = 0.25, respectively). Regarding the outcome of grades in assessments, it was noted that students who used AI tend to get grades greater than or equal to 80%, significantly (p = 0.007). Two articles were at high risk of bias, and five were of concern. It is concluded that the use of AI has been shown to be effective in the performance of tasks by Dentistry students. It is recommended that new AI tools be tested, and studies with low risk of bias be developed, to reach more robust conclusions about the impact of AI on the performance of dental students.