What if healthcare becomes not a universal right but a personal upgrade? Asks Paul Armstrong

The healthcare industry is entering its most radical transformation since the antibiotic era – and possibly its most dangerous. AI is no longer just scanning X-rays or arranging appointments. Robotic systems are beginning to perform surgery with levels of precision humans cannot sustain. Generative models are designing drugs from molecular blueprints. Exoskeletons promising longer, stronger working lives are moving from prototype to pilot. Elon Musk talks about a future where AI surgeons operate endlessly and where machine reasoning supports every clinical decision. A future like that could expand access, prolong healthy years and reduce suffering – and a future like that could just as easily produce an Elysium-style divide where optimisation is reserved for the wealthy and the rest of society receives the algorithmic, or, gulp, genetic, minimum.

Healthcare providers, insurers and governments are betting heavily on the optimistic version. Predictive diagnostics can identify cancers years earlier. DeepMind’s AlphaFold has mapped protein structures that accelerate drug discovery timelines from decades to months. Machine learning triage systems process clinical data at speeds no human team could match. McKinsey estimates AI could deliver more than one trillion dollars of annual savings for global healthcare systems by 2030. The NHS is already investing in AI radiology tools and automated triage models in an attempt to ease chronic pressure. A healthier population and a lighter budget is an appealing equation for any administration.

A sharp divide emerges once deployment begins. Data shapes every AI medical system, and most datasets come from affluent, urban, Western patients. Minority groups and lower income communities appear under-represented in the very material that trains models. A review in Nature Medicine found that most commercial medical AI tools showed........

© City A.M.