Use of Artificial Intelligence in the real-world dermatology setting

This thesis investigates the use of Artificial Intelligence (AI) for skin cancer diagnosis. It explores how personality characteristics, cognitive styles and biases may impact adoption and safety of an AI diagnostic tool. An interventional trial found that the AI diagnosed skin lesions with high accuracy, comparable to dermatologists. The AI also improved the diagnostic accuracy of less experienced clinicians. An online reader study and survey demonstrated that clinicians have high confidence in the AI and perceive it to be useful. Clinicians who change their diagnoses more often after interacting with the AI have less dermatology experience and higher agreeableness.