@article{article_1982987, title={Artificial Intelligence and Dentistry: A Clinical Review and a Proposed CHAIR Framework}, journal={Aydın Dental Journal}, volume={12}, pages={147–159}, year={2026}, url={https://izlik.org/JA26KD82XA}, author={Şengül, Doğan}, keywords={Diş hekimliği, İş akışı entegrasyonu, Risk yönetimi, Sağlık teknolojilerinin benimsenmesi, Yapay zekâ}, abstract={<p> <strong>Objective: </strong> Artificial intelligence (AI) is entering dental practice through imaging software, screening initiatives and patient-facing chatbots, yet clinicians have limited concise guidance on how to evaluate these tools. This review summarises where AI now meets dentistry and offers a practical method for appraising a tool before chairside use.  </p> <p> <strong>Materials and Methods: </strong> We conducted a narrative review of peer-reviewed studies, regulatory texts and reporting guidelines, favouring work that was validated beyond its own development sample, deployed in real settings or explicit about bias and governance. The evidence is organised along the clinical pathway and paired with a proposed five-point checklist (the CHAIR framework) and a worked example.  </p> <p> <strong>Results: </strong> The strongest evidence lies in radiographic image analysis. Convolutional neural networks detect caries on bitewings, measure periodontal and periapical bone loss and segment teeth and bone on cone-beam computed tomography at a level comparable with experienced clinicians. Photograph-based models screen for oral cancer and its precursor lesions, while other systems assist cephalometric analysis, extraction planning and implant identification. Language models are starting to help with documentation and patient communication. Accuracy is highest on tightly defined tasks and usually declines when the scanner, patient group or disease distribution differs from the system’s training data; reliable testing in new clinical settings remains uncommon.  </p> <p> <strong>Conclusion: </strong> AI is most useful when framed as augmented intelligence: a support for the dentist’s judgement, not a substitute. Safe adoption depends on appraising each tool against the local setting, monitoring for bias and putting clear governance in place. </p>}, number={2}