Would you want a game-show-playing computer to be your doctor?

If you feel like your doctor never has enough time, believe it or not — these may just turn out to be the good old days.

Because the doctor’s visit of the future might not even involve him at all.

IBM is hoping its Watson computer, made famous in television’s Jeopardy, will soon be allowed to help doctors treat patients.

To do that, it turns out, the company needs to convince Congress that Watson isn’t really a “medical device,” but in some other technology category. One that doesn’t need to be proven safe and effective.

Now think about this for a minute. Sure, your doctor may not always be right. But can we really trust something that answered “What is Toronto?” in a category called “U.S. cities”?

And is Watson any better at coming up with answers than, say Google?

No really, according to Douglas Hofstadter, a scientist at Indiana University.

Hofstadter says that Watson, despite how clever and impressive it seemed on Jeopardy, is just a “text search algorithm” connected to a database. Which is exactly what Google is.

And, Hofstadter says, it doesn’t actually “understand” anything. In fact, it can neither read nor comprehend what it’s saying.

And that might be the real danger in allowing Watson into the exam room.

What if Big Pharma scientists got in charge of the programing?

Now IBM said it’s ready to spend a cool billion to make Watson into a product that it can sell to doctors. All it needs is for a bill to pass Congress that would give the FDA two years to figure out a way that Watson (and likely many “relatives” of his) can bypass any FDA control.

But as amazing as Watson seems to be, Hofstadter says that we shouldn’t lose sight of the mystery that makes people even more amazing. It’s called “the miracle of human thought.”

And that’s something a machine will never be able to demonstrate, no matter how well it did on Jeopardy.

Sources:

“This medical supercomputer isn’t a pacemaker, IBM tells Congress, NewsOK, January 29, 2015, newsok.com

“Why Watson and Siri are not real AI” William Herkewitz, Popular Mechanics, popularmechanics.com