Diagnostic accuracy of artificial intelligence for detecting dental caries in camera and smartphone photographs

Introduction: Artificial intelligence applied to intraoral photographs could extend caries detection beyond the dental chair, but its accuracy on camera and smartphone images has not been consolidated. Methods: A systematic review was conducted, including studies reporting the diagnostic accuracy of artificial intelligence for detecting dental caries on intraoral photographs taken with cameras or smartphones. Results: Seventeen primary studies published in 2025 and 2026 were included. Reported accuracy was moderate to high but varied by acquisition device and lesion severity: detection on camera-acquired photographs was generally more accurate than on smartphone photographs, and cavitated lesions were detected more accurately than initial lesions. General-purpose multimodal models underperformed dedicated convolutional models. Conclusions: Artificial intelligence detects dental caries on intraoral photographs with moderate to high accuracy, but the evidence is young and dominated by single-institution studies. Accuracy is higher for camera than smartphone images and for cavitated than initial lesions, and dedicated models outperform general-purpose ones.