An autonomous robot recently sewed up a sliced intestine in an anesthetized pig. With the sutures spaced evenly and the pig resting peacefully in the recovery suite, the surgery was a success, representing yet another breakthrough in artificial intelligence.
“Supervised Autonomy”
When the Soft Tissue Autonomous Robot (STAR) sewed up the aforementioned pig intestine, it was aided by a surgeon at key points in the procedure. Though many people fear the ramifications of increasingly intelligent machines, surgical robots represent one version of the ideal relationship between “autonomous” technology and people: the robots have the ability to act independently, but are always closely supervised by people.
Imagine doctors as micromanaging bosses and robots as new employees. The more training the robots receive, the more comfortable the doctor feels delegating different types of tasks to it.
Doctors have been training other types of health care professionals for years. As health care has become increasingly complex and the number of people responsible for providing it has grown, it was only a matter of time before robots took on a more central role.
Increasing the Supply of Health Care
Given that robots are diagnosing diseases, conducting medical research, and even exhibiting empathy, its tough to come up with tasks in health care that couldn’t conceivably be automated at some point in the future.
Doctors, nurses, and other health care professionals must undergo years of training before they’re able to provide care, often expending hundreds of thousands of dollars in the process. This creates an enormous barrier to entry and reduces the number of people able to become doctors or nurses. The result: labor makes up 54% of a typical hospital’s operating revenue.
If machines could perform some of the tasks that require the most training (e.g. surgery), this could further shift the role of the doctor from pure task-performer to supervisor/trainer. Moreover, it could reduce the number of people required to provide care. The typical hospital margin is only 2.2%, and all hospitals are required by law to provide “charity care” to those that cannot afford to pay. With such small margins, hospitals often struggle to provide enough charity care to satisfy the demand while still remaining financially solvent.
If artificially intelligent computer programs and machines eliminate even a small portion of a hospital’s labor expenses, they could drastically increase margins, freeing hospitals up to perform more charity care. And that is exactly what they would do. Almost all hospitals are not-for-profit entities, and healing sick patients is central to their mission.
Hospitals are only one setting in which people receive care, but artificial intelligence is likely to have an impact beyond the hospital, making it possible for higher acuity care to be performed in new types of settings. Overall, artificial intelligence has the potential to drastically increase the available supply of health care services.
A Universal Human Right
The demand for health care services, however, has always been—and will always remain— high. No matter how healthy we are, at some point in our lives, we all need to receive health care.
Right now, your ability to access high quality care plays a huge role in determining whether you live or die when you become seriously sick. Those living in low income countries are 13 times more likely than those in high income countries to die before the age of five. Even in the US, people separated by only a few miles can have up to 20 year differences in life expectancy.
We all yearn for a society with true equality of opportunity, but it can never happen when the same disease is easily treatable for one person but deadly for another.
The promise of artificial intelligence is to change this. Whether it meets this promise remains to be seen, but that fact that AI is merely in its infancy leaves much room for optimism.
Intersections 
Great article, Maor. Really liked this.
LikeLike