Theoretical physicist, Centre de Physique Théorique, Aix-Marseille University; author, The First Scientist: Anaximander and His Legacy

There is big confusion about thinking machines, because two questions are always mixed up. Question no. 1 is, How close to thinking are the machines we’ve built or soon will build? The answer is easy: immensely far. The gap between our best computers and the brain of a child is the gap between a drop of water and the Pacific Ocean. Differences are in performance, structure, function, and more. Any maundering about how to deal with thinking machines is premature, to say the least.

Question no. 2 is whether building a thinking machine is possible at all. I’ve never really understood this question. Of course it’s possible. Why shouldn’t it be? Anybody who thinks it’s impossible must believe in things like the existence of extranatural entities, transcendental realities, black magic, or the like—must have failed to digest the ABCs of naturalism: We humans are natural creatures of a natural world. It’s not hard to build a thinking machine—all it takes is a few minutes of boy and girl and then a few months of girl letting things happen. That we haven’t found other, more technological ways yet is accidental. If the right combination of chemicals can produce thinking and emotions—and it does, the proof being ourselves—then surely there should be many analogous mechanisms for doing the same.

The confusion stems from a mistake. We tend to forget that entities composed of many things behave differently from those composed of few things. Take a Ferrari, or a supercomputer. Nobody doubts that they’re just (suitably arranged) piles of metal and other materials, without black magic. But we hardly imagine that a (nonarranged) pile of the same materials could run like a Ferrari or predict the weather like a supercomputer. Similarly, we generally fail to see that a pile of materials could (if suitably arranged) discourse like Einstein or sing like Janis Joplin. But it might, the proofs being Einstein and Joplin. Of course, that takes quite a bit of arranging and quite a lot of details, and a thinking machine will take a lot of arranging and details. This is why it’s so hard to build one besides the boy-girl way.

We have a view of natural reality that’s too simplistic, and this is the origin of the confusion. The world is more or less just a large collection of particles arranged in various ways. This is just a fact. But if we then conceive of the world as we conceive of an amorphous bunch of atoms, we fail to understand the world, because the virtually unlimited combinatorics of these atoms is so rich as to include stones, water, clouds, trees, galaxies, light rays, the colors of sunsets, the smiles of girls in spring, and the immense, black, starry night—as well as our emotions and our thinking about all of this. It’s hard to conceive of these phenomena in terms of atomic combinatorics—not because some black magic intervenes from outside nature but because the thinking machines that are ourselves are too limited in their thinking capacities.

In the unlikely event that our civilization lasts long enough and develops enough technology to build (in a way different from the boy-girl way) something that thinks and feels as we do, we will confront these new natural creatures just as we’ve always done, just as Europeans and Native Americans confronted one another, or as we confront a previously unknown species of animal—with varying mixtures of cruelty, egoism, empathy, curiosity, and respect. Because this is what we all are: natural creatures of a natural world.