Robot performs first fully autonomous surgical task – a blend of AI and precision?
By willowt // 2025-07-10
Copy
 
  • AI-powered SRT-H surgical robot achieves 100% accuracy in autonomously performing a critical step of gallbladder surgery on pig cadavers.
  • Trained via mimicry of human surgeons in video recordings, the robot adapts to anatomical variability and unforeseen challenges.
  • Researchers compare the advance to self-driving cars, signaling a shift from teleoperated tools to independent medical systems.
  • The milestone reduces reliance on human surgeons, with potential to address healthcare disparities and surgeon shortages.
  • Next steps include expanding the AI’s training for full surgeries and eventual clinical use in humans.
On Wednesday, researchers at Johns Hopkins University announced a groundbreaking achievement: a fully autonomous surgical robot, SRT-H, performed a complex, 17-step phase of gallbladder removal on eight pig cadavers without human guidance. This marks the first time an AI-driven robotic system has successfully executed such a detailed surgical operation—once exclusive to human surgeons—including adapting to unexpected anatomical variations and self-correcting during procedures. The milestone, reported in Science Robotics, underscores the rapid evolution of medical robotics and raises hopes for more accessible, consistent healthcare in underserved regions. While experts caution against overestimating timelines for human clinical use, the experiment demonstrates progress toward robots handling routine surgical tasks, allowing surgeons to focus on decision-making while improving outcomes.

The experiment: Training AI to think like a surgeon

The SRT-H robot learned its surgical skills through language-guided imitation learning, leveraging videos of human surgeons performing the same procedure on pigs. By analyzing visual cues and written task descriptions, the AI framework—similar to models powering ChatGPT—developed the ability to identify ducts and arteries, strategically place clips and sever tissues with tools like scissors. During testing, the robot faced realistic challenges: varied pig organ structures, shifted starting positions and simulated blood pools that altered visual cues. Despite these obstacles, it achieved 100% procedural accuracy, though its 5.3-minute average completion time lagged behind the 4-minute pace of expert surgeons. “This is a proof of concept that it’s possible,” said Axel Krieger, the Johns Hopkins roboticist leading the research. “Autonomous systems can now tackle the unpredictability of **actual patient care.”

A leap beyond remote controls: Why this matters now

Current surgical robots, like Intuitive Surgical’s $10 billion Da Vinci System, operate as sophisticated tools—controlled entirely by humans. Krieger’s team, however, has pushed toward artificial intelligence capable of autonomy. This distinction mirrors the shift from guided missiles to self-navigating drones. Unlike 2022’s STAR robot, which required pre-marked tissues and rigid plans, SRT-H adapts dynamically. Lead author Ji Woong “Brian” Kim emphasized the breakthrough’s practicality: “AI can now be reliable enough for surgical autonomy—once a distant goal, now within reach.” The development comes amid a surge in surgical demand. Over 2.7 million robotic procedures occurred in 2024, yet 40% of the global population lacks access to surgical care. Autonomous robots, argues Krieger, could fill gaps in underserved areas and reduce human error—80,000 preventable U.S. surgery-related deaths annually. Still, challenges remain. “Surgeons aren’t obsolete yet,” noted Johns Hopkins surgeon Jeff Jopling. “SRT-H is still slower, and full surgery testing on humans is years ahead.”

Moving forward: Balancing potential and pragmatism

The research, funded by ARPA-H and the National Institutes of Health, will now broaden to teach SRT-H complete procedures, such as kidney removal or colorectal surgeries. “This is modular progress,” said Jopling, noting that mastering phases like gallbladder dissection allows incremental training toward broader autonomy. Public trust and regulation pose hurdles. David Rescott, a Baird analyst, predicts ethical debates over AI medical decision-making without guarantees of infallibility. Meanwhile, the system’s real-time voice-response feature—enabling surgeons to correct the robot—suggests a hybrid future where humans and machines collaborate. “Think of it as an AI surgical fellow,” said Ji Woong Kim. “It learns from expertise but executes with machine precision.”

Promise and caution in the operating room

The Johns Hopkins experiment illustrates the waning boundaries between human skill and algorithmic precision in healthcare. While autonomous robots are not replacing surgeons overnight, they could become indispensable tools for repetitive or high-stakes tasks, much like calculators revolutionized mathematicians’ work. For patients, this could mean faster recoveries, reduced costs and equitable access to care. Yet, as with any technological leap, humility is key: “SRT-H shows we can,” said Krieger, “but we must proceed with rigorous testing and transparency.” The era of autonomous medicine is approaching—not as a replacement, but as a partner on the journey toward humanity’s healthiest future. Sources for this article include: Reuters.com Hub.JHU.edu CTVnews.ca
Copy