AI-driven Dentistry
Rosana Farjaminejad, Mahsa Jalali, Alexander Garcia-Godoy, Franklin Garcia-Godoy, Abdolreza JamilianBackground
Artificial intelligence (AI) is increasingly being applied in healthcare and dentistry. Advances in machine learning, deep learning, natural language processing, reinforcement learning, and generative AI have introduced new approaches to dental-data analysis and clinical decision support.
Objective
This review summarizes current applications of AI across dental specialties and evaluates their potential contributions to diagnosis, treatment planning, risk prediction, and individualized patient care.
Methods
Published evidence concerning AI applications in oral and maxillofacial radiology, orthodontics, prosthodontics, periodontology, endodontics, pediatric dentistry, oral pathology, restorative dentistry, implantology, oral surgery, and dental public health was narratively reviewed. Particular attention was given to the AI models used, their clinical applications, reported performance, and barriers to clinical translation.
Results
AI-based approaches have been extensively investigated for interpreting dental images and detecting caries, periapical lesions, oral pathology, and alveolar bone loss. Convolutional neural networks and transformer-based models have shown promising performance in radiographic analysis. Hybrid and ensemble models can combine radiographic, clinical, behavioral, and biological information for risk assessment. AI has also been used to simulate orthodontic tooth movement and craniofacial development, support digital prosthesis design and esthetic planning, predict periodontal disease progression, estimate dental age, assess pediatric behavioral patterns, and predict early childhood caries. However, many systems rely on small or nondiverse datasets, and independent external and prospective clinical validation remains limited.
Conclusion
AI has the potential to improve diagnostic consistency, treatment planning, risk prediction, and personalized dental care. Future research should prioritize transparent and interpretable models, diverse multicenter datasets, prospective clinical validation, privacy-preserving collaborative learning, and integration with existing clinical workflows. AI should augment rather than replace clinicians’ professional judgment.