Rendering a Verdict
in which our hero ponders generative AI
Dear Readers,
Like most artists, I have strong feelings about Generative AI. The ability for the average person to type a prompt into their computer and get a custom-made image or bit of text is understandably intoxicating. I’ve even played with the stupid things, just to see how they work. And yet, once you dig into the hidden costs of this “magic,” it gets harder and harder to see the use of Generative AI as anything other than morally reprehensible.
First, most (if not all) Generative AI apps are trained on stolen creative work. Companies feed these machines with images and books they scrape from the Internet, typically without the original creator’s permission—because, of course, that would cost the companies money and they hate that. They only want to make money. It’s everyone else who should be spending it.
Second, the data centers used to generate all this stuff place a tremendous strain on the environment of our already overtaxed world.
But, with all of that said, I do wonder if we AI haters are lumping too many things together into one KFC’s Famous Bowl full of slop.
This week on YouTube, the folks at Royalty Now decided to share where they use Generative AI in their process. And that really got me thinking. I think what Royalty Now does—creating facial reconstructions of historical figures—is artful. I can tell how much thought and skill goes into the creation of each portrait. Do I change my whole opinion of them now that I know they use AI in some small parts of their workflow?
I’m curious what you think. If someone uses Generative Fill in Adobe Photoshop to add trees to an otherwise barren-looking landscape they’ve painted, perhaps because they’re on a deadline, does that make them any less of an artist than the person who digs around a site like Unsplash or Pexels for a free-to-use photo of trees and photobashes the new element in themselves?
We all take shortcuts sometimes, right? And if the shortcuts free up our brains to do more difficult, creative work, aren’t they worth it in the end?
If that same artist is using AI to slot the trees in so that they can focus on painting the complicated character in the foreground to the best of their ability, do we really want to judge them the same way we judge a person who prompts the AI to create an illustration of someone’s dirty feet and then passes it off as their own work?
I don’t know. I’m just thinking out loud here. I still think most uses of Generative AI are deplorable and/or lazy. But the underpaid, overworked artist who uses these tools to help with the more mundane busy work involved in creation? Who am I to judge them?
Yours,
Chris
I think it's important to cut through the marketing enthusiasm and techno-evangelism that is propelling "generative AI" to the forefront of seemingly every aspect of American culture right now. It's not truly generative, as you point out, being perhaps more accurately named "regenerative AI" given the manner in which it blends and processes the data it has ingested into the intellectual equivalent of the "pink slime" from which Chicken McNuggets are formed before being battered and deep-fried.
My approach to AI - and to technological change in general - has long been shaped by the work of two of America's great cultural critics, Wendell Berry and Neil Postman.
In 1987 Berry wrote an essay called "Why I Am Not Going to Buy a Computer" - you can find it here: https://classes.matthewjbrown.net/teaching-files/philtech/berry-computer.pdf - that includes his proposed rules for the adoption of a new technology.
In 1992 Postman wrote a book called Technopoly: The Surrender of Culture to Technology, and in 1998 he gave a speech called "Five Things We Need To Know About Technological Change" - available here: https://student.cs.uwaterloo.ca/~cs492/papers/neil-postman--five-things.html - that is an excellent capsule overview of his position in the book.
I don't think that AI is beneficial for our society; I've learned to be deeply suspicious whenever someone extols the convenience of something or tells me that it will revolutionize the way something is done without any cost to the doer. There's always a cost, and I fear (not without reason) that the cost of AI will be the loss of our free will.