Healthcare

Healthcare presents a mixed AI displacement picture. Clinical roles - physicians, nurses, physiotherapists - are strongly protected by licensing requirements, physical interaction, and the high stakes of medical decision-making. However, the sector's enormous administrative burden is highly susceptible to automation.

Medical coding, billing, prior authorisation, transcription, and scheduling represent a large fraction of US healthcare employment. AI systems can handle these tasks with high accuracy and are already being deployed at scale. KLAS Research estimates that AI-assisted clinical documentation can save clinicians 2+ hours per day - which simultaneously increases physician productivity and reduces demand for medical scribes and transcriptionists.

Radiology and pathology are the clinical specialties most often cited as AI-vulnerable, with AI systems matching radiologist accuracy on specific tasks like diabetic retinopathy screening and certain cancer detections. However, these tools are currently being deployed to augment rather than replace radiologists, given the liability environment and the breadth of what radiologists actually do.

Long-term care, mental health, and complex chronic disease management are highly resistant to automation. The sector faces a genuine worker shortage rather than surplus in most clinical areas, suggesting AI will primarily drive efficiency gains rather than mass displacement.