We're shitting on AI for all the wrong reasons

Generative AI is a symptom of a larger problem—not the disease itself.

Two quick notes before we get into it:

1. This is Creativity Under Capitalism’s first new issue since going on summer break in June. Thanks to everyone who not only understood the necessity of breaks in avoiding burnout, but continued sharing the newsletter when there weren’t even new issues to share. Your support means the world, and I’m looking forward to making summer break an ongoing practice throughout this newsletter’s lifetime.

2. I link to several pieces of my own writing (published on ExtremeTech.com, my main freelance outlet) in this issue. AI is something I think, learn, and write about every week, thanks to what I do for a living, and I’ve therefore covered many facets of its existence in the past. ExtremeTech doesn’t earn anything special when you click these links; in fact, my editor doesn’t even know I’m including them. I’ve purely linked to these pieces because they offer important background insights, the sources for which I’ve already verified.

---

As ChatGPT, DALL-E, Midjourney, and built-in text and image generators become more ubiquitous, so too do everyday internet users’ opinions about them. This is fantastic: everyone deserves to have a say about something that is rapidly changing the internet as we know it, and an open dialogue about generative AI is necessary for facilitating informed decisions about those programs’ uses. But of all the perspectives I’ve read on social media, seen in magazines, or heard from friends, one in particular rubs me the wrong way. It sounds a little bit like this:

I get the sentiment behind statements like this one, I really do. Why are people so adamant about creating, advertising, and using AI programs that “write,” make music, and generate visual “art,” if that’s what humans are best at? Surely our species isn’t made to launder clothes and wash dishes until the end of time, right?

But I worry that by pushing the narrative that AI is doing the “wrong” things, we’re ignoring the reason generative programs like ChatGPT, and DALL-E exist in the first place. To start, non-generative AI is helping humans take care of busywork that would otherwise take days, months, or years to perform—we just don’t hear about it often because the mainstream news focuses on sensationalist (read: despair-inducing) AI headlines. Some AI-powered programs are indeed helping doctors diagnose lung cancer early and plucking trash from our oceans and harbors. These technologies have the power to save lives, mitigate pollution, and, theoretically, give us a little more free time with which to create art.

Will we actually see that free time, though? Perhaps not. The world’s wealth gap is growing, and the middle class is disappearing. People are increasingly turning to second jobs to afford housing, food, healthcare, and other essentials. Yes, we’ve automated much of the work we once did by hand, thanks to manufacturing robots, home appliances, computers, and other technologies. (In fact, it’s laundry machines and dishwashers that allow you to turn away from your laundry and dishes for more than a couple hours each week.) But that automation has simply allowed the world’s richest people to make more money as the rest of us grovel for a living via other means. The working class doesn’t get to see the rewards of time-saving measures and reduced labor needs; executives, holding companies, and shareholders do.

It only takes one look at all the layoffs resulting from AI-driven automation in the fast food, sales, software development, and gaming industries to see that the disease is capitalism; generative AI is merely a symptom. In a hyper-competitive, “dog-eat-dog” economy, the optimization or elimination of one form of labor merely demands that another take its place as the folks at the top grow richer. Whether that optimization or elimination comes from AI or another technology doesn’t matter. This cycle will continue until the United States and other wealthy countries pivot toward an economic model that disincentivizes speed, separates survival from labor, and levels the playing field so that a handful of out-of-touch billionaires can’t decide what the internet looks like. (We can dream.)

This doesn’t mean generative AI is innocent. As James Worth recently pointed out in an incredibly relatable note, conversations about AI should focus a little less on whether AI-generated text or images are “good”—a subjective term that the world will never truly agree on—and a lot more on how horrible AI is for the environment. The servers that power ChatGPT consume over half a million kilowatts of electricity every day, which is enough to power 180,000 US households. Those servers require constant cooling, which means they also consume water: holding just one “conversation” with ChatGPT uses just under the equivalent of one Dasani water bottle. If you thought golf courses were bad, look at the hundreds of tons of water wasted every day via ChatGPT’s 200 million daily requests.

This environmental havoc doesn’t matter a whole lot to the aforementioned out-of-touch billionaires, who are accustomed to chasing revenue and “disruption” at all Earthly costs. It only matters to those of us who stand to face some of generative AI’s myriad consequences. It is the working class that already disproportionately encounters the effects of human-exacerbated climate change, and it’s the working class that will scramble to find new jobs when generative AI renders their life’s work redundant.

It follows that the only path toward a world with limited AI-related harm is one that also leads to limited AI-related profit for those on top. And halfway through 2024, it’s difficult to imagine whether this path will be an easy one to forge. Technology companies might be foaming at the mouth to incorporate AI into their products, but these tools—namely chatbots, the most popular form of generative AI—aren’t nearly as popular with everyday people as they’re believed to be. Some technology researchers at prestigious universities believe AI will only boost productivity by 0.5%, while financial experts at legacy firms think generative AI tools aren’t built to solve the complex problems they’d need to solve in order to be worthwhile. (You can read about these assertions, and others, in a report preserved on the Wayback Machine.) In terms of reducing AI-related harm, these are encouraging projections, and because MIT and Goldman Sachs are the types of institutions tech founders and billionaires will actually listen to, they could signal a shift in the long-term adoption of generative AI.

Still, it’s hard to tell whether generative AI will go the way of cryptocurrency and NFTs—internet fads that crashed, burned, and became a part of cringey tech bro culture almost as quickly as they appeared. In the meantime, it’s imperative that we frame discussions around generative AI within the right context: that generative AI is a tool of exploitation that cheats the working class out of its future while rewarding those on top.

What’s been inspiring me lately:

✰ Once There Were Wolves, a darkly beautiful novel about trauma, the rewilding exploited landscapes, and the gray area between right and wrong. I’d love to read more stories that explore how we determine whether someone is “good,” “bad,” or deserving of redemption—if you have any recommendations, please let me know!

✰ Paper Jam + Print, a collaborative print studio in Phoenix, AZ. I got to visit during downtown Phoenix’s First Friday art walk last week, and the vast number of paper type, paper color, book size, and binding options made me want to create and print anything I could. I highly recommend reaching out to them if you want to print a zine, small-run magazine, booklet, or flyer while supporting a small business.

✰ Bad Animal, a beautiful bookshop/restaurant/wine bar in Santa Cruz, CA. Not only was this a gorgeous space filled with new and used books, wine galore, and the smell of some indulgent dinner, but a person who worked there told me Bad Animal intentionally prices its books lower than other online or brick-and-mortar retailers. Why? “Everything else in this world is so expensive—you should at least be able to buy a book on the cheap.” Swoon.