It's one of the big stories of our time — so much so that the White House has said it will remake our society.
On Wednesday, I attended an Intelligence Squared debate at Manhattan's 92nd Street Y that made me see artificial intelligence in a whole new way.
The topic: "Don't Trust The Promise Of Artificial Intelligence." The debaters: Andrew Keen, Jaron Lanier, James Hughes, and Martine Rothblatt.
Lanier — who pioneered virtual reality — stressed that we need to divide "artificial intelligence" into two different things.
There's "the engineering and the science on the one hand," he said, "and then on the other, the storytelling about it, the narrative that we have about it, the fantasy life of it — perhaps the religion of it. These are two distinct things. It doesn't mean one is good and one is bad, but they're just different sorts of beasts."
So there's news about an algorithm beating a human in an ancient, infinitely complex game. Then there are promises that you'll upload your mind or that your personality will live on as a chatbot after you die. An algorithm is research, to follow Lanier's logic; a prediction of technologically enabled salvation is more like religion.
To Lanier, when we ask "What is the promise of an area of research?" the only answer is that we fundamentally don't know.
"It's research," he said. "We just observed gravity waves for the first time. Does that mean we'll suddenly have anti-gravity devices? Well, you know, maybe someday. We have absolutely no clue what we're going to discover."
Lanier emphasizes that we don't really know that a thought is. Science can't describe it. We've found "collections of neurons that seem to active at certain times," but we still don't really know what thinking — and by extension, consciousness or the mind — actually is. The idea of becoming immortal by uploading your mind is more a profession of faith than a scientific description.
So when we talk about "artificial intelligence," we should be clear to say if we're talking about the ideology that's grown around it — or the research itself.