Robot Driver vs. Robot Surgeon: Why is one easier to trust?

With more than 46 companies currently investing in R&D in the autonomous vehicles space, conversations have centered on how this innovation will disrupt transportation and spurred a lot of excitement about just how soon we might be able to just sit back and relax while our robot drives us around. Despite several security concerns related to self-driving car hackers, less than half of Americans have reservations about self-driving cars. Meanwhile, as incredible milestones have been reached in robotics surgery to improve outcomes and reduce costs, the road to winning over patients and doctors when it comes to autonomous surgery proves to be much more challenging.

This sentiment is exemplified in the story of J&J’s Sedasys, a machine which replaced the need for an anesthesiologist to be present during certain medical procedures, but then was taken off the market after low sales. Whether the decision was driven more by patient safety concerns or by opposition from the American Society of Anesthesiologists (based in fears of the machine making anesthesiologists obsolete), the message from the public seemed clear: it was difficult and uncomfortable to fathom a machine replacing human expertise in an operating room.

But why does this skepticism not translate as widely to driver-less cars? For most people, the thought of a car without a steering wheel getting you from point A to point B without a driver seems much more comforting than a robotic arm performing a procedure without a doctor in the room. The former is labeled futuristic, while the latter seems like a mistake away from a horror film. Perhaps it is because we think of surgery as an intricate, invasive task that needs to be handled with precision and consistency by highly-skilled doctors. However, in surgical experiments, Smart Tissue Autonomous Robots (STAR) have been shown to have more precision and accuracy than human doctors when performing the same tasks, and they even outperformed human doctors with regard to consistency and outcomes during a fully autonomous soft-tissue surgery on a pig. These types of surgeries are incredibly complex and indicate a significant advance in the application of smart technologies — such as image recognition, pressure sensors, and feedback loops — in reducing both the physical and decision burden on surgeons and, ultimately, avoiding human error.

These are the same technologies that support the argument for why self-driving cars are safer than human-driven cars, but for some reason, we as a society more readily accept this assertion than we do the safety and efficacy of robot surgeons. Furthermore, when considering both the sterile, isolated environment of an operating room and the relatively contained environment of a human body, it can be puzzling to understand why we think the outside world could be more predictable and more easily learned by artificial intelligence. Of course, each human body can be vastly different, but one could make the argument that there are an infinitely greater number of unique factors that a self-driving car might encounter out in the world that could cause a threat to safety (versus what a robot surgeon might encounter).

There is no question that the innovation ongoing in the autonomous vehicle space is indeed impressive — however, I think it’s fair to question why we don’t similarly appraise the future potential of the autonomous surgery advancements that are built on the same technology. Likely, it has been the incrementally increasing level of semi-automation in cars (i.e. cruise control, parking assist) that has gotten us to a point of comfort that allows us to more easily fathom complete driving automation in the future. When it comes to surgical procedures, commercialization and exposure has been slower, thus, patients and doctors alike are still much more comfortable with supervised, incremental autonomy and are hesitant to move beyond it as quickly. Of course, issues of security and ethics plague the level of autonomy exercised by machines in both situations — but perhaps they seem more visceral and affect a population perceived to be more vulnerable when imagined in an operating room. In summary, the smart technology and artificial intelligence applications to healthcare are promising — but whether they will gain enough momentum to actually disrupt the system and meaningfully impact cost, access and patient outcomes might depend more on convincing the public of the sophistication of the technology than the sophistication of the technology itself.