Autonomous Robots Gain Surgical Skills through Video Observation, Enabling Independent Procedures
The rise of artificial intelligence is making its mark in the medical field, with AI-driven visit summaries and patient condition analysis becoming commonplace. Now, novel research suggests that AI training methods akin to those employed for chatbot models like ChatGPT could be utilized to teach surgical robots to operate independently.
Scientists from John Hopkins University and Stanford University developed a training method using video recordings of human-controlled robotic arms performing surgical tasks. By learning from their actions, the researchers believe they can decrease the need to program each specific movement required for a procedure. As reported in the Washington Post:**
These robots mastered manipulating needles, tying knots, and stitching wounds without human intervention. Moreover, the autonomous robots went beyond mere imitation, self-correcting mistakes like picking up a dropped needle without any prompts. Scientists are now transitioning to the subsequent stage of this project: integrating various skills for surgeries conducted on animal cadavers.
Robotics have been a fixture in operation theaters for several years. For instance, the "surgery on a grape" meme from 2018 showcased robotic arms aiding in surgeries by delivering enhanced precision. Around 876,000 robot-assisted surgeries occurred in 2020. Robotic tools can access areas and perform tasks within the body that a surgeon's hand cannot reach, and they are resistant to tremors. Precise, thin instruments can also avoid damaging nerves. However, robotic procedures are typically directed by a surgeon using a controller. The surgeon ultimately has control.
The skepticism surrounding more autonomous robots stems from the fact that AI models like ChatGPT do not truly understand the concepts they interact with but instead mimic past experiences. This raises concerns regarding the infinite variety of pathologies in the unpredictable human body. What if an AI has not encountered a specific scenario before, or fails to respond appropriately in the event of complications during surgery?
Autonomous robots meant for surgical procedures would necessitate approval from the Food and Drug Administration. In contrast, AI is used to summarize patient visits or make suggestions without FDA endorsement, as the physician is responsible for verifying the information. This opens up the possibility of AI bots providing incorrect information or hallucinating facts. How frequently will an overworked physician blindly accept AI-generated information without thorough examination?
Recent reports suggest that Israeli soldiers rely on AI to identify potential threats without scrutinizing the information meticulously. Consequences ensued when poorly trained soldiers attacked humans based solely on AI predictions, such as when the only verification required was that the target was male. Beware of complacency and failure to remain vigilant in critical situations.
Healthcare is aDomain with significant consequences—far more substantial than the consumer market. If an email summary is incorrect in Gmail, it is not a major issue. However, an AI system misdiagnosing a medical condition or making a surgical error can have severe repercussions. Who would then be held accountable for any mishaps? The director of robotic surgery at the University of Miami offered his perspective on this matter:
"The stakes are extremely high," he stated, "because this affects life and death situations." Anatomy varies from patient to patient, and diseases manifest differently in individual hosts.
"I examine CT scans and MRIs and then perform surgery using robotic arms," explained Parekh. "If you want the robot to conduct the surgery on its own, it will require comprehending all imaging, reading CT scans and MRIs. In addition, robots will need to learn laparoscopic surgery, which utilizes small incisions."
The expectation that AI will ever be infallible is highly debatable, given that no technology is perfect. Despite interesting theoretical potential, the repercussions of a botched surgical procedure conducted by an autonomous robot could be catastrophic. Who would bear responsibility for such mishaps, and who would lose their medical license? Humans may not be infallible, but at least patients have the comfort of knowing they have undergone rigorous training and can be held accountable in the event of errors. AI is a simplified replica of humans, prone to unpredictable behavior, and lacks an ethical compass.
If physician burnout is a contributing factor (as suggested by researchers), perhaps the underlying issues causing this scarcity should receive attention instead. There have been numerous reports indicating a severe physician shortage in the U.S., with projections suggesting a deficit of 10,000 to 20,000 surgeons by 2036, according to the American Association of Medical Colleges.
In the future of surgical technology, the integration of artificial-intelligence training methods similar to those used for chatbot models like ChatGPT could potentially allow surgical robots to operate independently, reducing the need for specialized programming for each procedure (based on the text). Furthermore, as tech advancements in artificial-intelligence continue to evolve, these autonomous robots could revolutionize the medical field, making surgeries more precise and efficient (speculating from the context).
