AI is a threat to artists, but not because an AI can actually create art.
I wrote a post about ChatGPT a while back. I had a look at what AI can do, and speculated about the consequences. We’re now three months down the line, and the shape of the disruption is slowly starting to emerge.
AI startups are popping out of the woodwork. Large companies are vying for a position on this new lucrative market. And those whose jobs get disrupted by AI are left holding the short end of the stick as their livelihood is repurposed to make big tech richer.
Of course, I can hear some of you thinking ‘you Luddite’.
And Luddites are — in fact — quite relevant to the AI discussion. Time for a small detour.
Luddites
Back in the nineteenth century, the industrial revolution happened. For English textile workers, this meant that companies started to use machinery that put them out of jobs. These textile workers united and formed opposition groups, calling themselves Luddites, after Ned Ludd, a legendary weaver. The Luddites destroyed the new machinery and demanded skilled textile workers keep their jobs.
The UK government intervened and dispatched some 12,000 troops to deal with the Luddites. Over time, the term ‘luddite’ has become a pejorative way to describe people resisting technological changes.
So, applying this to AI, am I opposing technological change? Am I trying to stops the wheels of progress to the detriment of humanity?
Well, in fact, no, that’s not my intention. I think we can benefit from AI, and that it can make humanity better. But, that doesn’t make any and all applications of the technology good. Quite the contrary.
Distribution of wealth
As I wrote above, the Luddites made demands of the companies they worked for. Those demands were not ‘remove these machines’. No, they demanded minimum wages, using taxes to pay for pensions, and regulations for minimal working conditions. Company owners were creating dismal factories where cheap laborers spent their days operating textile machines. The Luddites wanted to protect the less fortunate.
The company owners refused to negotiate and instead pushed the government to send the military to take care of the Luddites. The reputation of Luddites as standing in the way of progress was undeserved. They didn’t want to turn back the clock, they wanted better living conditions. Corporate elites framed it differently, and had them beaten down. And got away with it.
This same specter haunts much of our recent technological progress. AI is the newest in a series of ‘disruptive’ technologies. These technologies have indeed improved our lives, but they’ve also been used to move wealth from the middle class to the a small elite. Uber, Airbnb, and social media have pulled the rug out from under certain professions. Making big tech richer, while making hard-working people poorer, and working under increasingly horrible working conditions.
If you follow the flood of AI news, you’ll see the shape of things to come. Corporations will start replacing costly professional artists, writers, and marketeers with AI. Instead, workers will add the finishing touches to whatever the AI vomits up, if the AI vomit can’t be used as is — errors are the consumer’s problem.
The chances of earning a livelihood as a writer or artist will shrink further. And the money that frees up? It goes into the pockets of corporations creating and using the AIs, and their shareholders.
Regurgitation
A big problem with AI is that it’s not really ‘intelligent’. It’s mathematical trickery to do predictive generation of text or images. And if it’s not intelligent, it’s not really creative. Instead, an AI is fed large datasets of images or texts and then tuned with more data to be able to recreate variations and extrapolations of what was fed into it.
There are already discussions about copyright on those datasets. It seems unfair that a big tech company gets to use the art of thousands (if not millions) of artists for a pittance, then make millions off of it. As I wrote, AI is a vessel to redistribute wealth.
But fixing this copyright issue will only do so much. There is several thousand years worth of art out there, and a lot of it isn’t copyrighted. If you combine all the copyright-less art with the freebies and public domain work floating around the internet, I think you’ll be able to create a pretty good AI.
And that’s the rub. You can regurgitate the public domain of art with an AI, and might successfully take a large chunk of the market for writing and illustrating. But you’re only regurgitating existing art. AI-generated art is the Soylent Green of art. It’s stagnant.
Art is a reflection of our culture. The stories we tell and the illustrations we create tells those who come after us what moved us. So what does it say about us if we replace our art with remixed works of our past?
Reductionism
Again, when the above point comes up, there are always people who argue that human artists are also trained using the work of others, and they also don’t pay extra for that. So why is AI different?
On the face of it, this argument might seem to have merit.
If you look deeper, though, it’s extremely reductionist. If you applied this to painting you could say ‘painting a picture-by-numbers is really just the same as what Rembrandt did; it both involves brushes and paint’.
Yes, humans and AI both train using large sets of data and lots of practice. But an AI is a stupid algorithm. A human brings intelligence to the equation. A human uses art deliberately to say something about the world and life. And that makes all the difference.
Because AI is a trick. A nice trick, but a trick nonetheless. If you ask ChatGPT to tell you the meaning of life, it will tell you an amalgamation of what other people wrote on the subject. It will not think about the meaning of life. It will not set up a logical argument and try to teach you that. Instead, it steals that argument from writers who already did the thinking.
Meaning
The result of the two developments above is less artists and more regurgitated art. All in the name of enriching the rich. Is that progress? Am I a Luddite to want something different?
I love writing, and drawing, and painting. I have an almost pathological need to create things. And most of us do. As Chuck Wendig recently wrote:
“The human experience is one where we hope to be free to make art and tell stories and sing songs, so if artificial intelligence is doing that part, too, then what’s the fucking point of it all? When the robots make the music and the humans are cleaning up the dog shit — WTF?”
– Chuck Wendig
That’s the core of the Luddites’ problems, and the core of our current problems. The Luddites did not want to turn back the clock, they just wanted a decent life. And that’s what writers and artists want too. That’s what I want. I’d love to use AI, but not if the cost is taking meaning away from us and pushing droves of people into poverty and hard labor.
I had a discussion once with a colleague who argued that he always downloaded everything illegally. That you shouldn’t ask money for books, or for music, because that was a hobby and you should do those for free. I’m still stunned by this, because it’s basically saying that you should only be paid for a job if you hate it. That’s not how the work should work.
Smoke and mirrors
Interestingly, a lot of people working in AI are crowing about the dangers of the technology. The CEO of OpenAI (the company behind ChatGPT) has talked about the dangers. Elon Musk signed an open letter about the dangers of AI.
Are they on the Luddite route, hoping to save the middle class?
Nope. They seem to be purposely directing attention away from the humanitarian issues, and focusing on abstract dangers of the technology. The same CEO of OpenAI said OpenAI would leave the EU if they started to regulate the technology to strictly. And Elon Musk used to fund OpenAI and registered a new start-up just before signing that open letter.
But if we’re talking about the dangers of AI in military drones and nuclear war, we’re not talking about pushing writers and marketeers into poverty to make big tech richer.
In short
I think we can do a lot of great things with AI.
First though, we should stop calling it AI. It’s not intelligent.
And second, we should ensure we use this technology to make life better for everyone, not just funnel more money from the poor to the rich. We should not just use it to suck the joy out of life.
Heck, while we’re at it, we should do this with the other ‘advancements’ and ‘disruptions’ of the past few decades.
