This Is Not Inevitable
Image Source: https://www.popularmechanics.com/culture/movies/a28496339/rutger-hauer-death/
Agency. What a word. It captures our ability to affect what happens in our lives. Free will, the cosmic X button. Some call it the ultimate gift from a god.
And that was all Roy Batty and his fellow Replicants—artificial humans—wanted in Blade Runner.
A chance to employ their agency. To determine the course of their existence without someone’s thumb on the brakes. Which, in their case, was a predetermined and short lifespan.
And the thumb belonged to humans.
My fingers are working on keys so my thoughts can be shared. I‘m enacting agency. But I also do so just months after a set of infinite fingers, flowing from much greater hands, threw a shudder through anyone who arranges words for meaning:
Fuck, you’re saying, the last thing we need is another essay on that.
I agree!
I’ve read so many of them (daily), not to mention watched the news segments (hourly?), that I’m almost numb. And I wouldn’t have said anything if it didn’t seem like they were all missing a key point. Running right by the start of the story.
This is not inevitable.
“Inevitable” is the sense I’ve gotten from the wave after wave of excited prognostications and doomsday foretelling. Especially for writers. That a darker future, where we’re either blown out of work or adapt to coexist, at best, is simply going to happen. If you’re a Sandman fan, it’s almost as if we’re all reading Destiny’s book with the same agency-less paralysis: This happens now.
The problem of ChatGPT, of course, is that it’s doing something, from drafting passable ad copy to amusing poetry to not-all-that-bad screenplays, which we thought machines would never be able to do. Creators were supposed to be safe. But maybe they are not, according to Derek Thompson, who’s been doing some canny reporting at The Atlantic.
ChatGPT converses, as the “chat” implies. But what it really does, what we do when we talk, is convey meaning through words. How? Using the whole web, it responds to a question by searching like a search engine, analyzing fast as a microchip, and spitting out, in summary, what is most-probably a good answer. Or joke. Or advice. Or story. And evidently probability is the key element, or the probability of a word being the best one to use next in an emerging sentence. As Stephen Wolfram explains, it’s writing reduced to numbers and statistics—which is scary and enthralling and worthy of study, as he notes. The output, however, can be unpredictable, or just increasingly worse over time, like a Xerox copy of a copy, as Ted Chiang laid out. But whatever its batting average, what ChatGPT really does is write. Fast. And with access to a library so expansive even a fantasy novel would struggle to do it justice.
Which is causing panic among writers, not to mention really putting a wrench in publishing. That includes gumming up some sci-fi mags like Clarkesworld due to the spike in (spammy) volume. You don’t have to really write now, you just have to program. Same in the art world thanks to the other generative AI programs many are ready to gun down. So much for craft.
But ChapGPT is the special devil of anyone who constructs prose as a key part of their job. Like reporters. Teachers. Even lawyers.
See? It all feels inevitable, right? Progress arrives like a hurricane. Best we pray.
But it should be recalled that it is also a choice. All of it. To make AI like ChatGPT. To use it. To trust it.
We choose, with our agency. From the engineers to the CEOs to the venture capitalists, we are quickly selecting this future. And though few of the articles mention it, we seem to be heading into it at the dead run of a blinded horse.
But we can’t blame the cosmos. We can’t blame the creator. And I do not believe in fate; it absolves us of too much responsibility. We, collectively, are reality’s jockeys. If not its fattened and drunk owners.
I once imagined that before we all woke up on any given day, that the world simply didn’t exist. And in so many ways, it doesn’t. Commerce doesn’t. Politics don’t. Hate does not. Neither does love.
Sure, it’s foolish to think that a future absolutely shot-through and shaped by AI won’t happen. Equally, standing on the rails of progress is idiotic. If you want to believe in something reliable, choose probability. This train is running, coming in ways and a pace we can’t fully appreciate. And while the outcomes won’t be what we fear or hope, not totally, it is almost certainly going to occur. (Check out more of Derek Thompson’s best scenario imagining.) Is occurring.
I recently asked someone special if they would watch a movie or show scripted entirely by AI.
Their response?
“Sure, if it’s good.”
And that’s the truth. It’s also the essence of anything that’s been outmoded. A simple calculation: If it is good—if it is better, easier than what came before—we’ll use it. Because one of our primary drives, maybe the primary drive in the privileged world, is to use our agency to improve our individual lives.
As writers and editors, do we accept ChatGPT submissions? Do we use it for drafting, from technical exposition to ideas for period-specific dialogue? Do we signify when something has been enhanced? Do we pay more for “human-created” content? Do we roll out awards not for best novel but best “prompt engineer”? All of this should be wrestled with. And will be, I bet.
But it is not inevitable.
At each step change, each juncture, we have choices both as makers and consumers. We still build the future.
At least, of course, until we build machines, different kinds of life, that can do it better. That can make it good. Like Roy, in those last 42 words.
If we do that.
The choice is ours.