University of California San Diego researchers reported on July 8, 2026, that a surgeon remotely controlled a humanoid robot during two laparoscopic gallbladder operations on pigs. The peer-reviewed Nature paper describes the work as an in vivo feasibility study using teleoperation, benchtop tests and dry-laboratory comparisons.

The robot did not choose the surgical plan or act independently. A senior surgeon operated it from a console, and human clinicians carried out most bedside assistance in both procedures.

What the humanoid system did

The researchers built the system around the G1, a general-purpose humanoid made by China-based Unitree Robotics, and adapted it to hold wristed laparoscopic instruments. The console sent the surgeon's hand commands to the robot over a network, while a stereo camera supplied the operative view. The robot then moved the instruments through ports in the animal's abdomen.

The full study reports that attending surgeons and clinical fellows placed the abdominal ports before the robot was positioned, and a human assistant managed most camera control, tissue retraction, exposure and instrument adjustments. A second humanoid briefly held the camera and retracted tissue during the first case, but it did not replace the bedside team.

A clinician sits at a console used to control surgical equipment remotely (illustrative image)

Both operations followed the established steps of laparoscopic cholecystectomy, or gallbladder removal. The robot-compatible equipment did not include a clip applier, so clinicians clipped the cystic duct and tied off the cystic artery with conventional laparoscopic methods before robotic work resumed.

The 32-minute figure covers console time

The first case recorded 56 minutes 15 seconds of active console control. The second recorded 31 minutes 59 seconds, commonly rounded to 32 minutes. These were not total procedure times: the study separated periods when the surgeon actively moved the humanoid from pauses for camera cleaning, instrument changes, recalibration and repositioning.

The study's workflow table records eight robot deployments in the first case and four in the second, with major pauses lasting more than three minutes occurring twice and three times, respectively. The second operation also involved minor bile leakage and bleeding from the liver bed, which the team managed with suction and electrocautery. Neither case required conversion to conventional laparoscopy or open surgery.

How it differs from the da Vinci system

The humanoid and Intuitive Surgical's da Vinci platform share one basic feature: a surgeon controls the instruments from a console. Intuitive Surgical's system guide says every surgical maneuver remains under the surgeon's direct control and that the system translates hand movements into instrument movements.

The hardware is different. Da Vinci is a purpose-built surgical platform with dedicated arms, instruments, imaging and operating-room procedures. The research system starts with a general-purpose humanoid, then adds mounts, control software and laparoscopic equipment.

Articulated robotic arms operate in a clean technical setting (illustrative image)

The experiment does not establish that the humanoid is clinically equivalent to da Vinci. In a dry-laboratory peg-transfer task involving 13 medically trained participants, the da Vinci Xi produced faster average completion times, fewer errors and higher scores than the humanoid. The comparison tested a training exercise, not patient outcomes.

Why this is not autonomous surgery

Teleoperation transmits a person's decisions and movements to a machine. Autonomy requires the system to make at least some decisions or carry out parts of a task without continuous manual control. The pig operations tested the former, even though the platform may support more automated functions in later research.

A 2024 systematic review screened US regulatory databases and classified 49 surgical robots cleared by the US Food and Drug Administration from 2015 through 2023 on a five-level autonomy scale. It placed 86 percent at Level 1, robot assistance, and 6 percent at Level 3, conditional autonomy. The review found no Level 4 or Level 5 systems in that set.

A humanoid robot stands in a technology laboratory (illustrative image)

That review and the humanoid experiment cover different groups. The review examined surgical devices already cleared in the United States, while the humanoid paper reports a preclinical research platform tested in animals. Its results do not establish clearance, approval or clinical availability in any country.

Barriers before human use

The research team identified control latency, limited reachable workspace and sensitivity to calibration drift compared with established surgical systems. During the animal procedures, clinicians also encountered repeated repositioning, reduced strength and intermittent overheating. Those problems raised the workload for the surgeon and bedside team.

Sterility is another unresolved issue. The researchers covered the robot's arms with gloves, but said this setup did not reproduce the sterile workflow required for human surgery. Commercial humanoid components cannot generally be sterilized in an autoclave, and any protective drape would also have to preserve sensor performance and calibration.

UC San Diego's research release says the robots needed recalibration several times and that the team is still working on communication delay for longer-distance operation. The university presented remote communities and difficult field settings as possible future uses, not current clinical deployments.

APPI News could not verify a public regulatory filing or a timetable for human trials at the time of writing. Any future authorization and availability would vary by country.

Frequently asked questions

Did the humanoid robot operate autonomously?
No. A surgeon controlled the instruments from a remote console throughout the robotic portions of both procedures. Human clinicians also placed ports, performed steps that lacked compatible robotic tools and provided most bedside assistance.

Did a humanoid robot complete surgery in 32 minutes?
The second case logged 31 minutes 59 seconds of active console control. That measurement excluded pauses and should not be read as the total time spent in the operating room.

Was the system tested on people?
No. The paper reports two operations on pigs in a surgical simulation center. It does not report a human trial or routine clinical use.

Is the humanoid system equivalent to da Vinci?
No such equivalence has been demonstrated. Both are controlled from a surgeon's console, but they use different hardware, and the established da Vinci Xi outperformed the humanoid in the study's dry-laboratory peg-transfer comparison.