Robotic Surgeon Flawlessly Removes Pig Gallbladder Without Human Control

BALTIMORE — A surgical robot has successfully removed pig gallbladders in a lab setting without any hands-on help during the actual surgery steps, completing each operation with a 100% success rate. The new system, developed by researchers at Johns Hopkins University, handled all the critical parts of the procedure by itself, even correcting its own mistakes along the way.

Although a human operator still had to reload surgical clips and switch tools when needed (like a nurse would during a real surgery), the robot made all decisions for the surgical steps on its own, from spotting the right parts of the anatomy to placing surgical clips and making precise cuts. The team says this is the first time a robot has handled an entire surgical procedure, step by step, without direct control. A paper on the achievement was published in Science Robotics.

Teaching Robots to Think Like Surgeons

The secret behind the robot’s success is its two-part artificial intelligence system, which works like an experienced surgeon guiding a medical trainee in the operating room.

The first AI component acts like a lead surgeon: every three seconds, it looks at video from surgical cameras and decides what should happen next, issuing simple commands like “clip the left tube” or “move the right arm higher.” The second AI acts like the surgeon’s hands, turning those instructions into smooth, precise movements 30 times a second.

To train the system, the researchers recorded 17 hours of real gallbladder removal practice on pig organs taken from butcher shops and used for medical training. They captured more than 16,000 examples of the steps involved, including common mistakes and how to fix them, so that the robot could learn how to adjust on its own when something goes wrong.

How the Robot Surgeon Did

When tested, the robot completed all 17 steps needed to remove a gallbladder, including grabbing tissue, placing clips to seal off ducts and arteries, and cutting them. Each procedure took about five minutes on average, not counting time for the human assistant to reload clips and swap tools.

The robot proved it could adapt when things didn’t go exactly as planned. In some tests, the researchers deliberately put it in tricky positions, such as having to grab tissue at an odd angle. The system figured out how to fix the problem by itself. On average, it made about six self-corrections per surgery.

When compared to a simpler version of the AI that didn’t have this layered approach or built-in corrections, the new system was far more reliable. The basic model only succeeded about a third of the time, while the full version got it right every time.

“The motion of the robot felt very human and it felt like some living being was controlling it,” lead author Ji Woong “Brian” Kim, a former postdoctoral researcher at Johns Hopkins who’s now with Stanford University, tells StudyFinds. “This made sense — the robot was trained to imitate human motions, but still, its ability to self-correct mistakes and perform intelligent behavior was quite exhilarating to watch.”

Looking Ahead: How Robotics Might Help Patients

While these tests used pig organs in a lab, not live animals or humans, the breakthrough brings us closer to real surgical robots that can take on bigger roles in operating rooms. In the United States alone, surgeons remove over 700,000 gallbladders every year. Smart robots like this could help hospitals handle routine surgeries, especially in places where there aren’t enough trained surgeons.

“Just as surgical residents often master different parts of an operation at different rates, this work illustrates the promise of developing autonomous robotic systems in a similarly modular and progressive manner,” said co-author Jeff Jopling, a Johns Hopkins surgeon.

Importantly, these systems aren’t designed to replace surgeons entirely. Instead, they could take care of straightforward, repetitive steps, freeing up human surgeons to focus on complex parts of an operation or unexpected problems. Because the robot understands simple spoken instructions, surgeons could step in to guide it if needed, just like training a resident.

“This advancement moves us from robots that can execute specific surgical tasks to robots that truly understand surgical procedures,” said medical roboticist Axel Krieger. “This is a critical distinction that brings us significantly closer to clinically viable autonomous surgical systems that can work in the messy, unpredictable reality of actual patient care.”

Of course, plenty of hurdles remain before this kind of system is used on real patients. The robot still needs to be tested inside the body, where it would have to work around bleeding, moving organs, and tight spaces. The cameras it uses today are also too big for standard surgical tools, so they’ll need to be made smaller.

Patient safety will always come first. The team behind this research says future versions could check their own confidence and ask a human surgeon for help when they’re unsure. They might even be able to run a virtual test of each step before doing it for real. For now, this study shows that robots can do more than single tasks, they’re learning to handle entire surgeries from start to finish.