With Artificial Intelligence (AI) adoption in dentistry growing, a new study in Nature warns that clinical oversight remains crucial when integrating AI-driven diagnostics and treatment planning into practice.
Aga Khan University researchers reviewed 44 studies on AI in dentistry, identifying key risks, including misdiagnoses, transparency issues, and inconsistent adoption across specialties.
Case study: AI misidentification in caries detection
One example from the study involved a 12-year-old patient whose dental X-ray was analyzed by AI software. The system flagged a cavity on a lower molar, but closer examination revealed it was an artifact from a recently applied fissure sealant. Without human verification, such errors could lead to unnecessary treatments, highlighting the need for clinician oversight.
Transparency and real-world validation concerns
The study found that AI companies are not always forthcoming about their technology’s accuracy. Of several contacted for validation, only one responded, raising concerns about transparency in AI-driven diagnostics.
While AI tools are widely used, few independent studies evaluate their real-world performance, making it difficult for clinicians to assess their reliability in everyday practice.
AI adoption varies by specialty
AI has gained traction in orthodontics due to its role in precision-based treatment planning. However, its adoption remains limited in prosthodontics and periodontics, despite their need for accuracy in diagnosis and treatment execution. Researchers call for broader AI research to ensure equitable advancements across all dental disciplines.
Market growth and investment in AI
Despite these concerns, investment in AI-powered dental solutions is accelerating. VideaHealth, an AI-driven diagnostics company, recently secured $40 million in funding to expand into orthodontics, endodontics, and pathology. Meanwhile, market projections estimate the digital dentistry sector could grow from $5.4 billion in 2023 to nearly $20 billion by 2034.
However, researchers caution that rapid AI expansion must be matched with rigorous validation and regulatory oversight, including the incorporation of new patient-centered outcome metrics, such as morbidity. Meanwhile, the researcher did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.