A humanoid robot has removed a gallbladder, and for once the milestone is substance rather than spectacle. Researchers at the University of California San Diego reported this week that a general-purpose humanoid, nicknamed "Surgie," performed a laparoscopic gallbladder removal, the keyhole procedure surgeons do through a few small incisions. In a second experiment, two of the robots worked side by side to complete another. Both operations were carried out on non-primate mammals, not humans, and were framed explicitly as proof of concept. The work was published in the journal Nature under the plain title "In vivo feasibility study of humanoid robots in surgery."
What sets this apart from the robotic surgery already common in hospitals is the word humanoid. Systems like the da Vinci have assisted operations for two decades, but they are large, single-purpose machines built to do one class of task. Surgie is a general humanoid, the same body plan being pitched for factories and warehouses, pointed instead at an operating table. The UC San Diego team argues that versatility is the whole point: one adaptable machine that could handle many different procedures rather than a room full of specialized rigs.
The pitch is cost and reach. "It's a fraction of the cost and it takes a fraction of the space in an operating room," said Shanglei Liu, a surgeon on the team who teleoperated the robot. Michael Yip, a senior author, framed it as an access problem rather than a gadget: remotely operated or autonomous humanoids could bring surgery to places that currently have none, from rural hospitals to battlefields and, he suggested, eventually space. Many communities simply lack the staff to keep a surgical team running, he noted, and patients go untreated as a result.
The caveats matter as much as the achievement. Surgie did not operate on its own; a trained surgeon directed every motion through teleoperation. What was proven is that a humanoid body can execute a real surgical task under human control, not that it can reason its way through one. Autonomy is the stated goal, but it stays a goal. And this class of hardware has an uneven public record: the general-purpose humanoids now appearing everywhere are still prone to clumsiness and odd behavior once they leave tightly controlled settings. Trusting one to hold a scalpel unsupervised is a long way off.
Still, the direction is worth watching closely. The strongest argument for humanoid robots has always been that the human world is already built for human bodies, so a machine shaped like us can slot into hospitals, homes, and factories without redesigning them first. Surgery is about the most demanding test of that idea imaginable, and clearing even a proof-of-concept bar moves the conversation from party-trick demonstrations toward something with real stakes. The open question now is not whether a humanoid can perform surgery, but how much of a surgeon's judgment can safely be handed over, and how slowly that handover should happen.