Otiose

 Friends keep trying to creep me out about artificial intelligence even more than I already am. They show me websites that will generate an endless number of variations of original artworks; software that will write article copy for you that has never existed before. The visual productions are uncanny. The written prose the AI generates is somewhat less impressive. The quality of the writing is low, and it gives me more the impression of a search engine cobbling together component parts of the internet. It does not give me the spine-tingling subjective sensation others describe of communing with an actual mind. 

I admit, though, that I do worry about jobs. As I've argued before, the AI probably will not displace the need for human cognitive labor. The output of these machines is often flawed, and will require human workers to check and correct it. I can't imagine all companies opting to fire their human editors and replace them with machines, at least not anytime soon. I can, however, foresee an epoch when copywriters will be expected to save time and increase our efficiency by first generating the copy through an AI, and then editing and massaging it from there. 

Would this be so terrible? There are certainly worse things—such as outright unemployment. But there would be a loss nonetheless. Of course, there is nothing to prevent a professional copywriter from editing AI content during work hours and writing their own original content the old-fashioned way at home. There might even be a subsidiary market for human-only literary output that lasts far into the future, just as retro analog technologies stage periodic come-backs as status symbols among hipsters. But we would be in something of the same position as contemporary glassblowers. A handful of us could survive as a curiosity; but as an industry we would be extinct. 

Artists, writers, and philosophers have not even begun to grapple with the ramifications of this. They are still reading Walter Benjamin, when the work of art has long since left the age of mechanical reproduction and entered that of digital production. Expect William Morris-type revivals of "human writing" analogous to the mid-nineteenth century vogue for handicrafts—as an antidote to mass production under conditions of industrialism—once the reality of what is happening sinks in. 

While some of this anxiety will be justified, we should keep it in perspective. As I said above, it will probably not mean the annihilation of all jobs, any more than any prior technological revolution pushed people permanently out of work. But we should be aware that the same thing that happened to the 19th century handloom operators could still happen to us. And the tragedy would not just be having to find new employment. It would be, as it was for the weavers, the fact that we spent decades learning and perfecting a particular human skill, only to find we can no longer make any extrinsic use of it. 

There is a moment in William Kennedy's novel Ironweed, when the hobo protagonist encounters an old friend tending bar who used to be a radio star. The hobo himself played baseball as a younger man, before his downfall. Listening to the bartender deploy his famous singing voice, and thinking over both their fates, the hobo reflects on "the irrelevance of talent." He wonders: "How does someone get this good and why doesn't it mean anything?"

This is partly what we are all going through in the face of the AI revolution. How could the many years we spent perfecting one ability—in our case, Milton's "one Talent which is death to hide"—not count for anything? We know that we can continue to perfect it, and that has to mean something in a cosmic sense. The barkeep could still sing beautifully, and bring beauty into the world, even if he could no longer do it for a living; the hobo was still a great ball player in his time, even if it had not left him with anything to take into retirement. But if it doesn't do anything other than exist for itself—then what is it good for? 

This is what a friend keeps telling me when I try to comfort him about his drawing. "I was getting so good at it," he says, "but why continue if it won't lead to anything—because a machine can now do it just as well?" 

The answer may not suffice, but it is probably the only one available to us in a universe governed by the law of mortality and finitude. Even if there were no machines, after all, we would still have to figure out how to be content with the fact that—beyond a certain time horizon—our works will not last. Even if they conferred upon us wealth, we could not take it to the grave; even if they assured us of a posthumous reputation, eventually enough generations would pass for us to be forgotten; and in enough eons still after that, whatever medium had been used to record our thoughts would wither under conditions of entropy. 

AI or no AI, therefore, we still have to figure out how to be content with the fact simply of having lived and created for its own sake. Even if it doesn't lead to anything else; even if it doesn't outlast the moment. Still, we say, that does not deprive it of its reality in the present. As Archibald MacLeish once wrote: "They also live [...] Who swerve and vanish in the river." (We're back to echoing Milton again.)