Application and performance of artificial intelligence in implant dentistry: An umbrella review

This umbrella review evaluates the applications and performance of AI in implant dentistry, comparing its effectiveness to human intelligence. The PICO question addressed was: “In implant dentistry, how does the performance of AI-driven approaches compare to standard references or conventional methods performed by human practitioners?” A comprehensive search was conducted across databases such as PubMed, Embase, Scopus, Web of Science, PROSPERO, Cochrane Library, and Google Scholar until August 2024. Two independent reviewers conducted the screening, data extraction, quality assessment, and certainty evaluation. After screening 12 studies included for qualitative analysis. The majority of studies utilized deep learning (DL) models, and some studies employed traditional machine learning (TML) or simple rule-based algorithms. Most systematic reviews found AI applications and performance promising when compared to human intelligence. However, several challenges were identified, particularly in AI’s accuracy in measuring bone width and height, detecting the inferior alveolar canal, treatment planning, and predicting osseointegration. Although AI shows promise in detecting anatomical landmarks (such as the maxillary sinus), identifying implant systems, and supporting clinical decisions, current models still face significant limitations and should not yet be considered as standalone tools capable of replacing human practitioners.