Judgment under uncertainty
97Deciding what to do when a patient's real anatomy does not match the imaging or the plan is an in-the-moment call no current AI system can make or is trusted to make.
Surgery looks like an obvious automation target from the outside — precise, repeatable, physical — and parts of it genuinely are moving toward machines. But the core of the job is not the cutting motion; it is deciding what to cut, on a patient whose body never looks exactly like the plan, and living with the legal and moral consequences of that decision.
Robotic platforms and AI-read imaging are real and growing fast, and they are changing how many surgical tasks get done. What they have not changed is who is responsible when a patient's actual anatomy, mid-operation, turns out not to match the scan — that decision, and the accountability behind it, still belongs to a human surgeon with a hand inside the body.
Very low
Share of the work a machine could do
A small share of surgical work — some suturing, imaging analysis, routine endoscopic screening and post-operative monitoring — is realistically automatable in the near term. The center of the job, deciding what to do under uncertainty and carrying responsibility for that decision, remains firmly outside what any current robot or AI system can do, or is legally permitted to do.
Scored from the tasks, not the job title. Lower is safer.
Jobs AI cannot take →What machines cannot take
Legal and moral accountability
95A named human surgeon signs the consent form and answers for the outcome; no jurisdiction has worked out how, or whether, to put a machine's name there instead.
Manual dexterity in unpredictable tissue
90Real tissue bleeds, scars and varies between patients in ways a controlled simulation does not; today's surgical robots are precise tools a human operates, not autonomous actors.
Crisis response and improvisation
88When something goes wrong on the table — sudden bleeding, a cardiac event — the ability to improvise beyond the plan in seconds is what actually saves the patient.
Reading the whole patient
78Weighing a patient's overall health, values and family circumstances when deciding whether to operate at all is often more consequential than any single technical step.
What they already take
Pre-operative planning & image analysis
68AI models increasingly flag tumor margins, measure anatomy and help plan surgical trajectories from CT and MRI scans faster, and sometimes more consistently, than a human reviewing alone.
Routine endoscopic screening
60AI-assisted polyp detection during colonoscopy, and increasingly automated scope navigation for simple diagnostic procedures, already outperform unaided human attention on some narrow detection tasks.
Standardized, repetitive suturing
45Robotic and AI-assisted suturing devices have been demonstrated for simple, repetitive wound closure in controlled settings, though not yet for complex or variable tissue.
Post-operative monitoring & documentation
70Continuous vitals tracking, early-warning deterioration algorithms and automatic operative-note drafting from recorded video and audio already offload real clerical and monitoring work from surgical teams.
How the work is changing
The surgeon becomes a console operator, sometimes
Robotic platforms like the da Vinci system shift some tasks from direct hand-in-body work to tremor-filtered, console-based control, changing the physical skill set even as the underlying judgment stays entirely human.
AI as a second pair of eyes, not a second surgeon
Real-time image-recognition tools now flag anatomy, tumor margins or instrument proximity to critical structures during an operation — a warning layer added to human attention, not a decision-maker replacing it.
Subspecialization keeps deepening
The generalist surgeon of the mid-20th century has given way to increasingly narrow subspecialists, a decades-long trend toward deeper expertise in smaller domains that AI-assisted tools are accelerating rather than reversing.
Remote and tele-mentored surgery
Low-latency robotic links and video let an expert guide, and occasionally operate on, a patient continents away, changing where surgical expertise physically needs to be rather than what expertise is needed.
New jobs branching off
Surgical robotics technician
Installing, maintaining and troubleshooting robotic surgical platforms inside hospital operating suites, a hybrid engineering-and-clinical role that barely existed before the 2000s.
Surgical AI/data specialist
Evaluating and integrating AI-based imaging, planning and monitoring tools into a hospital's actual surgical workflow, sitting at the boundary between clinical practice and software.
Simulation-based surgical educator
Running high-fidelity simulation labs — virtual-reality, haptic and cadaveric — that now carry a growing share of early technical training once learned solely on live patients.
Tele-surgery coordinator
Organizing and technically supporting remote guidance or robotic-assisted operations across distant sites, work with particular value for the regions the Lancet Commission identifies as most underserved by surgical care.
Outlook
The tasks most exposed to automation are the ones that can be fully specified in advance: a screening scope, a standardized closure, a scan a model has seen thousands of examples of before. The tasks most protected are the ones that cannot — what to do when the abdomen does not look like the scan, when a patient's blood pressure suddenly drops, when the textbook plan meets a body that did not read the textbook.
That split is unlikely to eliminate the job; it is more likely to keep concentrating value in judgment and accountability while automating the increasingly large share of technical execution that can be standardized. A surgeon in 2040 will very likely operate with more machine assistance than one today, and will still be the person whose name is on the outcome.
Nothing in the current trajectory of robotics or AI changes who is legally and morally responsible when a decision under uncertainty turns out to be wrong. Until that changes, if it ever does, the center of the job stays human.
Keep exploring
More in Medicine & Life
🩺
Physician
Diagnoses illness and manages health for years afterward through examination and evidence, not a single operation — medicine's generalist and long-term guide.
AI-resistant 67 🐾Veterinarian
Treats patients from goldfish to horses that cannot describe a symptom, and carries the legal power to end their suffering by euthanasia.
AI-resistant 78 🧠Psychiatrist
Diagnoses and treats mental illness through medication and therapy, one of medicine's only specialties with legal authority to detain a patient in crisis.
AI-resistant 73