Diagnostic Decision Support Systems (DDSS) are AI technologies that analyse patient data to suggest diagnoses, recommend investigations, or propose treatment options to healthcare providers. These systems augment clinical decision-making but typically require human oversight and final clinical judgment. In Australian healthcare insurance, DDSS create shared liability scenarios where both the system developer and the clinician using the system may bear responsibility for patient harm. Insurance coverage must address professional liability for clinicians who rely on DDSS recommendations, particularly where the AI suggests incorrect diagnoses or misses critical findings. Key considerations include whether the system has TGA approval as a medical device, the level of clinical validation, clear guidelines for appropriate use, and the clinician's duty to exercise independent judgment. Policies typically cover the healthcare provider's decision to accept or reject system recommendations rather than the algorithm's performance itself.

Diagnostic Decision Support Systems (DDSS) are AI technologies that analyse patient data to suggest diagnoses, recommend investigations, or propose treatment options to healthcare providers. These systems augment clinical decision-making but typically require human oversight and final clinical judgment. In Australian healthcare insurance, DDSS create shared liability scenarios where both the system developer and the clinician using the system may bear responsibility for patient harm. Insurance coverage must address professional liability for clinicians who rely on DDSS recommendations, particularly where the AI suggests incorrect diagnoses or misses critical findings. Key considerations include whether the system has TGA approval as a medical device, the level of clinical validation, clear guidelines for appropriate use, and the clinician's duty to exercise independent judgment. Policies typically cover the healthcare provider's decision to accept or reject system recommendations rather than the algorithm's performance itself.