“What is it about robots, that makes them so robotic?” asks a recent video promo for the film “Prometheus,” which is directed by Alien and Blade Runner director Ridley Scott. The semi-robotic voice of “David” in the clip continues: “At Weyland industries, it has long been our goal to create artificial intelligence almost indistinguishable from mankind itself.” Later, When asked what the android David thinks about, he replies by saying “almost anything.”
More than a half century ago, mathematician Alan Turing reflected on the potential of artificial intelligence, eventually reaching the conclusion that computers would one day have intelligence rivalling that of our own.
A paper he wrote titled Computing Machinery begins with Turing asking whether machines could think. As it turns out, thinking, is difficult to define, so Turing considered the possibility of an “imitation game,” in which a computer could model natural conversation so closely that it would be impossible to distinguish it from a human’s. A robot that has reached this point would be said to have passed the “Turing Test.”
A recent piece from Wired reported that artificial intelligence could be on the brink of passing the threshold of the Turing Test. And futurist Ray Kurzweil has optimistically predicted that computers will have intelligence comparable to humans’ by the 2020s.
So what does all of this have to do with medicine? A lot, potentially.
It remains unlikely that androids will replace physicians any time soon (or ever, as many argue), but AI is beginning to play an ever-more important role in medicine—especially after IBM’s Watson is now being employed to help advance the practice of medicine by empowering physicians. As the mountains of data generated from everything from medical journals to medical devices grow, sifting through it to glean insights is an increasingly important need.
In 2010, Google executive chairman Eric Schmidt in 2010 brought up the topic of computer technology in medicine while addressing Checklist Manifesto author Atul Gawande, MD. Schmidt explained how he envisioned that algorithms could be used to help doctors diagnose diseases:
My question has to do with the model of health care that we'll be facing in 5 or 10 years. It's pretty clear that we'll have personalized health records, and we'll have the equivalent of a UPC sticker with your medical history. So when you show up at the doctor with some set of symptoms, in my ideal world, what would happen is that the doctor would type in the symptoms he or she also observes, and it would be matched against the data in this repository. Then this knowledge engine would use best practices, and all the knowledge in the world to give physicians some sort of standardized guidance.
[...] As computer scientists, this is a platform database problem, and we do these very, very well, as a general rule. And it befuddles me why medicine hasn't organized itself around these platform opportunities
|This conversation between two chatbots seems more artificial than intelligent. But speech recognition technology is continuing to progress at a rapid clip.|
While building a search engine is an entirely different matter than treating patients, the use of computing technology still, obviously, holds great potential as medicine becomes more and more like an information technology.
It was only last year that IBM announced that its Watson technology would be used for medical applications. And improving medical care is goal that is vastly more noble than defeating human opponents in Jeopardy.
In many ways, Watson is well suited for this task. Among Watson's most impressive capabilities is the fact can process natural spoken human language. And it can currently process 100 million pages of data of medical literature per second. In medicine, its potential to help clinicians keep up with research breakthroughs and epidemiology are huge. Watson could also help clinicans with everything from improving efficiency to reducing the number of medical errors to helping them be more objective in clinical decision-making.
When it was first announced that Watson would be used for medical applications, opinions in the physician community were mixed. While some welcomed it, others saw it as misguided with some of them even fearing "a robot takeover of medicine — and the planet."
Which brings me to my last point: why does the word "creepy" so often accompany descriptions of artificial intelligence? This is a complex topic but it deals largely with our fear that computer technology will ultimately disempower us.
It also makes us question what makes us human; we have never before encountered anything that is capable of emulating human emotion or intellect.
“The best way to predict the future is to invent it.”
And then there is the whole "uncanny valley" theory that suggests that humanoids that come close, but don't succeed, in imitating humans repel us.
In any case, now would be a good time to determine how computing techology can be harnessed to advance medicine—for both patients and physicians. As computer scientist Alan Kay has said: "The best way to predict the future is to invent it."
Brian Buntz is the editor-at-large at UBM Canon's medical group. Follow him on Twitter at @brian_buntz.