The aim of this review is to assess patients' knowledge, attitudes, and perceptions (KAP) regarding artificial intelligence (AI) in dentistry, as well as to understand the main concerns associated with its implementation. The electronic search conducted in seven databases retrieved 4,005 articles, of which 12 were selected for full-text reading through a double-blind screening process, nine met the inclusion criteria, and another four articles were identified through a citation search. The Joanna Briggs Institute Critical Appraisal Checklist for Analytic Cross-Sectional Studies was used to assess risk of bias, GRADE was used to assess the quality of evidence, and a random-effects meta-analysis was used to quantitatively assess the pooled event rate and pooled mean KAP values using 95% confidence interval. The overall risk of bias of the studies was as follows: low (8), moderate (2), high (3), and most studies were classified as low quality evidence according to GRADE. High heterogeneity values were found across various outcomes, but most indicated the same tendencies toward acceptance or disagreement with the statements. However, for knowledge/awareness scores, some studies had very high values and others very low, indicating variance in knowledge across the populations analyzed. Few results with low heterogeneity were found. The results suggest that participants value the dentist's work and would welcome AI as a supportive tool. Concerns remain regarding the risk of overreliance on AI and issues of trust, ethics, and privacy. Further research on patient opinions regarding the use of AI in dentistry is encouraged. • Patients have a positive attitude towards AI as a support tool in dentistry when these technologies are supervised by dentists. • Ethical concerns, data privacy, liability, and the need for professional training and tool validation remain important issues for patients. • Education and dissemination of information about AI technologies in the dental field are still uneven across different populations.
Lima et al. (Tue,) studied this question.