Artificial Intelligence in Dermatology: Assessing Predictability in Clinical Diagnosis

The study assessed the Tibot AI application's diagnostic performance in dermatology with 400 patients, finding high accuracy for adnexal disorders (AUC 0.93–0.98), pigmentary disorders (AUC 0.88–0.94), and cutaneous tumors (AUC 0.87–0.95). However, it showed poor sensitivity for immunological disorders (31.3%) and cutaneous infestations (22.2%). The AI's accuracy improved with top-three predictions, indicating its potential as a diagnostic aid rather than a standalone tool. The study highlights the need for diverse training datasets and integration with clinical expertise to enhance AI performance, especially for complex disorders. Conducted at a single tertiary care center, the findings may have limited generalizability.