■ SAFE RISK ■ Healthcare
No. War-zone medicine is defined by everything AI cannot supply: hands inside a body cavity, a decision made in ninety seconds with no imaging, and a willingness to be in the building when the second strike lands.
“Operating in a bombed-out hospital with no power. AI can advise. Humans do it.”
Our AI replacement risk score — how we score jobs
A day here rarely resembles a day in a teaching hospital. You triage a mass-casualty influx in a corridor, sorting blast injuries by who can be saved with the four units of blood you actually have. You do damage-control surgery — stop the bleeding, control contamination, close later — on a table lit by headlamps because the generator is rationed to the oxygen concentrator. You improvise: chest drains from IV tubing, external fixators from whatever the logistics convoy managed to get past the checkpoint. Between cases you negotiate with armed men about who gets through the gate, and you write the security incident report that determines whether your team stays or evacuates.
Automation has genuinely arrived at the edges. Handheld ultrasound with guided-acquisition software helps non-specialists find free fluid in an abdomen. Telemedicine links a general surgeon in Idlib to a burns specialist in Amman. Supply forecasting, epidemiological tracking of cholera or measles outbreaks, and translation of consent conversations all benefit from machine tools, and drone resupply is real in a handful of theatres. Predictive triage scoring can rank casualties faster than a tired human.
What resists is nearly all of it. Robotic surgery needs a stable platform, reliable power, a sterile field, and a bioengineer down the hall — a shelled hospital offers none of those. Consent and ethics under fire, deciding to stop resuscitating a child because the oxygen is finite, holding a team together after a colleague is killed: these are moral acts performed by a person who can be held accountable. Our risk score of 8 reflects a job where technology is a useful passenger and never the driver.
Automatability: our editorial assessment of current and near-term AI capability
Resilient for the foreseeable future. Decision-support and remote consultation will keep improving through the 2030s, and a lone surgeon in a besieged town will increasingly have an expert in her ear. But nothing on any credible roadmap puts a machine in a collapsing hospital doing a laparotomy by torchlight. If anything, worsening conflict and climate displacement mean demand for these doctors rises faster than the supply of people willing to go.
Grimly, yes. Humanitarian medical organisations are chronically short of surgeons, anaesthetists and emergency physicians willing to deploy, and armed conflict plus climate-driven displacement are expanding the caseload rather than shrinking it. The constraints on this career are burnout, funding cycles and personal safety — not automation. Job security here is the least of anyone's worries.
Mostly logistics and diagnostics. Machine tools help forecast drug and blood consumption, flag disease outbreaks from clinic data, transcribe and translate patient encounters, and guide inexperienced hands through ultrasound windows. Telemedicine platforms with AI triage route cases to remote specialists. None of it operates, cannulates a shocked child, or decides who gets the last unit of packed cells.
Not usefully in the near term. Robotic platforms demand stable electricity, controlled sterile environments, high-bandwidth low-latency links for any remote element, and technicians for maintenance. Conflict settings supply the opposite of all four, and the equipment is a looting and targeting risk. Militaries research remote surgery for casualty evacuation, but the practical trajectory is better telementoring for human surgeons, not robot replacements.
Breadth over subspecialisation: damage-control surgery, obstetrics, paediatrics, anaesthesia and trauma resuscitation with minimal kit. Add practical competence in ultrasound, austere anaesthesia and infection control. Then the non-clinical half — security awareness, humanitarian negotiation, working through interpreters, and teaching. Fluency with tele-consultation tools is now genuinely useful, but it supplements clinical judgement rather than substituting for it.