The View from Here

We speak about the future as if we could see it, and we can't. All of us are born into a certain age, with a certain set of technologies and assumption about normality, and it's easy to see this as human nature. You can see it in people's complaints about AI. Much of the worry starts from the assumption that some capabilities have to stay human: if machines write, remember, do maths or reason, we are deprived of some important aspect of our nature. Those qualms may be justified, but they all take for granted that we already know what capabilities human intelligence must have if we're to have what we want from it in the long run.

Others have been worried about this before. In Plato's Phaedrus, Socrates relates the story of the Egyptian god Theuth who offers King Thamos a gift for memory. The king refuses, and says that writers who write things down will stop remembering and look wise and never truly be so. For him, that was a perfectly reasonable worry, since a good deal of a person's intellectual life was to do with their memory, but we only know his argument because Plato thought it worth writing down, and the writing eventually gave us a lot more than it took away. Now that memory can be "out of the head", we can build on each other's ideas over centuries instead of just one, and the fruits of science and law and philosophy can become things no one mind could ever contain. Tools keep doing this; a calculator means you don't do long division by hand, a computer obscures enormous complexity behind an icon, and science only works because we can ignore most of the details below, each time handing off a task and working one level up.

So then, maybe AI is the next step (or really the next big step). That might not make us smarter, and we might not want to accept all automation that becomes possible. But I would hesitate to rule these things out simply because they don't align with our current expectations for what is best for a human. We are currently standing at a very narrow window of human history. For generations, repeating or memorizing epic poems, navigating without a map - these were critically important skills; now they are less so; there are skills we see as defining of being human today that may disappear entirely. I am not saying we should not regulate AI, limit it, and occasionally say no. I am saying we should do so with the recognition that we don't really understand what human intelligence is yet.