Surgeons Remotely Controlled Humanoid Robots to Perform Live Animal Surgery — a World First

Two humanoid robots stood next to an operating table this week. They weren’t folding laundry or running a factory line. They were holding scalpels.

A team of engineers and surgeons at the University of California San Diego used two off-the-shelf Unitree G1 humanoid robots to perform gallbladder removal surgeries on live, large non-primate mammals. In one procedure, a surgeon worked alongside a single robot. In the other, a surgeon sat at a console and controlled both robots remotely — with no one touching the patient directly. The results were published July 8 in Nature.

The obvious question is why you’d use a general-purpose humanoid robot instead of a dedicated surgical system like the da Vinci. The answer is mobility and cost. A typical surgical robot weighs around 1,800 pounds (about 816 kg), requires a team of specialists to install, occupies a huge footprint in an operating room, and often needs the room to be physically modified to fit. Humanoid robots are smaller, lighter, and can be shipped to a rural clinic or a disaster zone without weeks of setup.

The robots were controlled through a console by surgeons who watched a monitor and moved the robot arms in real time. The research team — engineers from the Jacobs School of Engineering working alongside surgeons — designed the control system to translate the surgeon’s hand movements into precise robotic actions at the operating table.

But the paper is upfront about what went wrong. The robots required multiple recalibrations during each procedure, which pushed surgery times well beyond what a dedicated system would take. Signal latency from remote control also caused delays that need to be resolved before the approach could be used on humans.

This is the first time humanoid robots have been used for live animal surgery, and the fact that two different modes — collaborative and fully remote — both worked suggests the concept is viable even if the execution still has rough edges. The next challenge is making the robots fast enough and reliable enough for human patients. That means cutting calibration time drastically and eliminating the control lag.