By Jonny Lupsha, Wondrium Staff Writer
Ethical concerns about artificial intelligence (AI) plague as many industries as AI can affect. If machine learning reaches a significant point, how much responsibility—and blame—should AI take on?
Tiny robots called “microswimmers” may soon be controlled with laser guidance systems. These microswimmers are just 2.18 micrometers across, meaning that 40 of them could stand side by side on the end of a human hair. Currently, they must be manually controlled by magnets or electric pulses, but scientists are considering testing them for use in delivering medicine directly to specific locations in patients’ bodies.
In his video Ethical Health Care in the Age of AI, Dr. David Danks, Professor of Philosophy and Psychology at Carnegie Mellon University, discussed the benefits and risks of AI actually administering the care for various health care needs, such as autonomous AI treatments.
I, Dr. Robot
The question has been raised already in science fiction, but as sci-fi gets closer to being scientific fact, humanity will eventually have to ask itself: What happens if we try to remove human doctors and nurses from the health care equation completely, even if just for basic procedures?
“Some people might immediately object that treatments should never be automated—a human should always be ‘in the loop’ monitoring and watching how the treatment is provided,” Dr. Danks said. “But the reality is that there are already simple automated systems that can, for example, alter the flow of a drug into an IV, based on a patient’s condition.”
Systems like this don’t involve any particularly advanced or complicated AI, but according to Dr. Danks, they exemplify how machines and automation are already involved in the treatment process. Although few people would raise concerns there, automation in medicine doesn’t end with IVs.
“In 2016, a research team from the Children’s National Health System in Johns Hopkins University developed a robot that could perform limited soft tissue surgery,” Dr. Danks said. “Soft and slippery materials are normally very difficult for robots to manipulate, and of course the human body is filled with such materials. Even seemingly simple surgical tasks such as suturing or providing stitches turn out to be very complex because of the nature of the human body.”
Amazingly, this robot outperformed human surgeons. Meanwhile, other systems are being developed for AI to diagnose both physical and mental illnesses.
A Tragic Irony of Need
If AI systems and robotic surgeons could be developed and proven to be safe beyond the shadow of a doubt, the debate over health care being “a right or a privilege” could enter a whole new arena.
“Modern health care systems around the world face shortages of doctors, nurses, and other health care professionals,” Dr. Danks said. “These shortages are particularly acute in some of the neediest parts of the world—millions of people around the world have limited or perhaps no access to reliable, high-quality health care.”
AI and robotic treatment systems could provide these places with the health care they so desperately need, but only if the price is right. This can lead to a concern that the inequity of health care around the world could only worsen with the implementation of AI-based care. Uncommon cases that require abstract or creative solutions may also cause them trouble.
“AI and robotics systems are often better than humans on average or typical cases, but at the same time they’re often much worse on unusual or oddball cases,” Dr. Danks said. “Imagine a full AI diagnosis plus treatment system—we can even imagine that the system provides better-than-average care for most people—but that very same system will probably provide worse-than-average care for unusual cases.
“As with most medical advances, many people will be helped, but a few might be harmed.”
Although fully capable and widely implemented robotics systems of health care are years or even decades away, these questions will become more pressing as their technologies are realized.