Twenty-two dental experts (6 women, 16 men) participated in the study, all having more than two years of clinical experience and regularly detecting caries. The study involved using a dental AI decision-support system called dentalXrai Pro 1. 0. 4. The system automatically detects pathologies and restorations on bitewing images, providing color highlighting and generating written documentation. Participants viewed 20 randomly selected bitewing images, with half of them having AI support and half without. The task was to visually inspect the images and report any proximal caries detections and treatment decisions. The participants used the AI software when available and toggled the AI overlay on and off.

The study utilized eye tracking technology to record participants' gaze patterns during the task. Data collection resulted in 349 datasets for analysis. The study design aimed for convenience and controlled for potential fatigue or learning effects. The trial was not conducted on patients but on retrospectively sampled imagery material. Ethical approval was obtained, and written informed consent was collected from all participants.