Background colour

PREVIEW

Video

AssetID: 54919715

Headline: Surgical Robot Performs First Ever Autonomous Gall Bladder Surgery

Caption: Surgical Robot Performs First Ever Autonomous Gall Bladder Surgery. A surgical robot trained using videos of real procedures has successfully completed a key phase of a gallbladder removal on a lifelike model - without any human assistance. The robot, developed by researchers at Johns Hopkins University in the United States, responded to verbal instructions during the operation, learning and adapting much like a junior surgeon guided by a more experienced colleague. The robot demonstrated the poise and precision of a seasoned human surgeon, even when faced with the sorts of complications often encountered in real-life emergencies. The system, known as the Surgical Robot Transformer–Hierarchical (SRT-H), is built on the same machine learning architecture that powers AI tools such as ChatGPT. It can interact using natural language, responding to spoken instructions such as “grab the gallbladder head” or corrections like “move the left arm a bit to the left.” In contrast to earlier systems, which required specially marked tissues and tightly controlled environments, SRT-H can adjust to different anatomical features in real time, make independent decisions mid-operation, and correct its own mistakes. To prepare, the robot was shown videos of Johns Hopkins surgeons performing the procedure on pig cadavers. Each video was paired with captions explaining the steps. After training, the robot completed the surgery with 100% accuracy. Although it took more time than a human surgeon, the outcome was on par with expert performance.

Keywords: Science & Technology,Surgical,Robot,First Ever,Autonomous,Gall Bladder,Surgery,videos,real,procedures,gallbladder,removal,lifelike,model,human,assistance,Johns Hopkins University,United States,precision,surgeon,operation,real-life,emergencies,Surgical Robot Transformer–Hierarchical,SRT-H,instructions,AI,ChatGPT,tissues,anatomical,captions,training,accuracy,performance

PersonInImage: