The Problem With AI is Us

In an Age of Extinction, We’re Looking for Life in All the Wrong Places

Image Credit: Snowy

See that? That’s my little tiny dog Snowy. He’s not an artificial intelligence. He’s a real one. Let me sum this whole essay up, so you don’t even have to read it. If you’re obsessed with AI….get a dog. Because…

You can barely crack open a newspaper these days without Yet Another Gushing Story About AI. For example, today in the New York Times, Kevin Roose — who’s a good writer — goes…to…chat…with a bot…and walks away shaken to the core. He can’t sleep! Is it alive?! Does it love him?! Can an algorithm…

Replace us? This, after all, is what the fear of AI is really about. AI’s going to replace us. The human parts of us. It’s going to write our books and compose our songs. It’ll script our TV shows and write our films. Why, is there anything AI can’t do? But then…what about us…the poor, tiny, fragile humans?

Relax, Carl. It’s gonna be OK. I’m here to talk you through the promise and peril of AI. First, let’s begin with the peril. It’s not that AI’s going to replace us, because, well, it isn’t. Go ahead, have a listen to a song written by it. Enjoy that particular experience? OK, now imagine replacing all your actual friends with chatbots. Or your wife and kids. Or your dog. Here, go right ahead and read a story written by an AI. AI is no match for us humans, at what we excel at, which is being human, and there’s a reason for that, which I’ll come back to.

Nor is AI going to “evolve” much past this point — in fact, it’s a mature technology. It’s the culmination of decades of research — almost a century, at this point, into things called “neural nets,” which are basically gigantic correlation engines, simplistic models of how the brain was once thought to work, only we know now that it’s not remotely that simple. We don’t understand, really, on any serious level how exactly minds work, because they appear to work at levels way, way beyond us still, research wise — quantum, subatomic levels. AI is just…correlations. It’s not remotely at that level of interaction, and it’s never going to be, not even if we have quantum computers, because nothing appears to suggest minds are just associative algorithms, simply saying, hey “this thing is like this other thing!” If it were that simple, well, you’d hardly have Picasso or Jackson Pollock or Einstein. Minds are capable of creativity in a way that’s hard to even express, so dramatic and unforeseen is it. Intelligence — the real thing — is something deeper entirely.

So what is it? Well, first of all, let’s begin with the concept of “Artificial Intelligence.” Why does it have this grandiose title? Because it’s a Cartesian idea. I think, therefore I am. It’s the old flaw of the Enlightenment, writ large. Human beings were at the top of the Great Chain of Being — and some humans were more human than others — because of “intelligence.” And by intelligence in this context was meant analytical skill, basically, and not much more: human beings could pick things apart, reduce them to their constituent parts, and draw out causes and effects. We alone could do that, and so it was said that we alone were intelligent. Hence, a few centuries later, “Artificial Intelligence.”

Only we now know that all the above is wrong. Completely, utterly, totally, hopelessly wrong. Mind isn’t just analytical intelligence. There are many kinds of “intelligence,” from social to cultural to interpersonal to relational to creative and so forth. And even putting them that way is a mistake, because they aren’t the same concept. Those other forms of “intelligence” aren’t about analysis and prediction and correlation at all. Sure, it’s “intelligent” of me to predict that my lovely wife is going to be mad at me if I don’t do the dishes…but that’s not what makes us partners. Maybe you see my point.