The Robot in Your Operating Room Doesn't Operate. A Human Does, for Now

There’s a phrase that means less than it sounds like: “robotic surgery.”

It conjures a machine bending over a patient, making decisions, cutting on its own judgement. That is not what happens. In an operating room running the world’s most common surgical robot, the machine is not the surgeon. It’s the surgeon’s hands, relocated a few feet away, and it does not do a single thing the surgeon doesn’t tell it to, in real time.

Understanding that gap between the image and the reality is the key to understanding both how far medical robotics has come, and how far it still has to go.

What the da Vinci actually is

The dominant system, made by a company called Intuitive and named da Vinci, is everywhere now. In 2025 alone, surgeons performed about 3,153,000 procedures with it, up roughly 18% in a single year on the figure Intuitive reported in January 2026, and that number climbs every year. By any measure it’s one of the most successful robots ever built.

That dominance is only now being tested. In July 2026, Johnson and Johnson’s OTTAVA system won FDA authorisation for ten general surgery procedures, the first real crack in Intuitive’s long monopoly in soft-tissue robotic surgery.

But here’s how the da Vinci works. The surgeon sits at a console, sometimes across the room from the patient, looking into a high-definition 3D view of the surgical field. Their hands grip controls. As they move, the robot’s several arms (tipped with tiny instruments inside the patient’s body) mirror those movements exactly. It’s called teleoperation: remote human control, with no independent decision-making by the machine at all.

What the robot adds isn’t judgement. It’s superhuman steadiness and precision. It filters out the natural tremor in a human hand. It scales motion down, so a two-centimetre movement of the surgeon’s hand becomes a two-millimetre movement of the instrument. It bends and rotates in tight spaces a human wrist can’t reach, through incisions smaller than a fingertip. The result is less bleeding, smaller scars, faster recovery: real benefits for real patients. But the intelligence in the room is still entirely biological.

Why “the robot did it” is a useful fiction

If the human is doing all the deciding, why call it a robot at all? Because the machine changes what the human is capable of. A surgeon operating through da Vinci can do things (in dexterity, in precision, in access) that the same surgeon’s bare hands cannot. It’s less a robot in the sci-fi sense than a very sophisticated power tool: a translator that turns skilled human intention into steadier, smaller, more precise action than a body can manage alone.

That framing matters, because it tells you exactly what the da Vinci is not: it is not autonomous, not self-directed, and not (despite the marketing shimmer of the word “robot”) making medical choices. When something goes wrong on a da Vinci case, a person is deciding what to do about it, always.

It would be too neat, though, to say that of every surgical robot. One review of 49 systems cleared by the US Food and Drug Administration between 2015 and 2023 sorted them by how much they do on their own, and the picture is a spread rather than a rule. About 86% have no autonomy at all: the da Vinci’s category, a human hand behind every motion. But roughly 8% carry out individual tasks by themselves, and about 6% run with what the researchers call conditional autonomy, handling a stretch of work while a surgeon supervises. The machines that make no decisions are the overwhelming majority. They are not the whole field.

The line that’s starting to move

Which brings us to the futuristic part, and the reason this is worth watching closely.

Researchers have spent years developing a very different kind of machine: the Smart Tissue Autonomous Robot, or STAR, designed not to mirror a surgeon but to actually perform a step of surgery itself. The project began in 2011 under the engineer Axel Krieger at Children’s National Hospital in Washington DC, and travelled with him when he later moved to Johns Hopkins. Soft-tissue surgery is the hard case: unlike bone, soft tissue is slippery, deformable, and never twice the same, which is exactly why it had stayed a human monopoly.

STAR crossed real milestones. It performed autonomous procedures on the intestines of live animals (connecting soft, moving tissue with stitches) and in some tests produced results as good as or better than human surgeons doing the same task.

Then came a separate system, SRT-H, built at Johns Hopkins with Stanford and written up in Science Robotics in July 2025. Having learned in part by watching video of surgeries, it carried out a seventeen-step gallbladder removal from start to finish with nobody driving it, and got through all of them with a 100% success rate. Read the small print, though, and the frontier is closer than the headline suggests: the work was done on eight pig gallbladders on a bench, outside a living animal. No blood pressure, no breathing, no body moving around the instruments. This is still no longer mirroring a human. It is a machine executing surgical steps on its own, in the easiest possible version of the room.

The question we’ll have to answer

None of this means an autonomous robot will be operating on you next year. These are careful research milestones, in animals and on animal tissue, hemmed in by enormous questions of safety, regulation, and liability. The gap between “did it in a lab on a pig” and “approved to do it on your grandmother” is exactly the kind of long, unglamorous chasm that self-driving cars spent fifteen years crossing.

But the direction is unmistakable. Today’s surgical robot is a brilliant instrument with a human mind behind every motion. Tomorrow’s may make some of those motions itself, and then we’ll have to decide, as patients and as a society, how much of the most intimate thing one human can do to another we’re willing to hand to a machine.

For now, the reassuring truth holds: the robot in the operating room doesn’t operate. A person does. It’s the two words “for now” that should have your attention.