What AI Can Never Achieve in Art
In the second half of 2024, I remember the chatter around AI picking up. Writers weren’t sure what to think, but these interactive tools were also not fully rolled out. AI still seemed like a somewhat far-off fog. And besides, it was a presidential election year. I coped, in part, by starting to watch Christmas movies in October and listening to the Wicked soundtrack on repeat. Who had time to worry about AI?
By the summer of 2025, my entire Google suite had been overtaken by AI prompts asking if I would like help summarizing documents or replying to an email.
Um, no.
How do I turn this thing off?
Also over the summer, I added my name to the class action lawsuit that was recently settled against Anthropic, a company that has used pirated books to train their AI systems.
Around the same time, I listened to an interview with Sam Altman, founder of OpenAI, and it sort of infuriated me. This anger was already below the surface, but suddenly I had thoughts.
As a user, I’ve experimented with ChatGTP for my job per the request of a department director, and have talked with friends and family who are figuring out how it might be useful to them. I’m pretty neutral about project management. If you’re having a birthday party and want to have AI help you brainstorm all the tasks that need to happen, go for it! If you’re desperate for a new shampoo and conditioner and want to save hours of research, I’m sure AI can help with that. I’ve also used it to come up with soccer drills for my son’s team, which I’m coaching this fall—head coach, BTW—an essay for another time! To me, these kinds of tasks are low-stakes, not unlike me telling a personal assistant to do something and report back with ideas.
But as I listened to this interview, something deeper bothered me, beyond the privilege he was clearly speaking from.
Sam Altman’s enthusiasm for the future of AI is operating under the assumption that this is a reality everyone actually wants.
During said interview, Sam was asked a question about how, in five years, will we be able to tell the difference between something real and something fake online. A very important question! His answer sort of implied we’d figure it out as we went along, that there were some technical things we could employ, but he definitely shrugged this off like it didn’t actually matter.
What if I don’t want to look at a photograph generated by AI? What if I want to look at a photograph taken by another human being who had a real human experience the moment that photograph was taken? That’s what art is.
Sam seems to be co-creating a reality he believes is both inevitable and beneficial, and again, assuming a number of things: That we’ll fall in line and adapt. That we want to! That AI should be the default, the norm, and if you don’t embrace it you get left behind. Oh well. Certain sectors will be obliterated. Oh well. Our children and grandchildren will be so sad for us that we ever had to function without AI in our lives. (He actually said a version of this.)
When Anthropic took the contents of my books, it was both without consent or compensation. When Google’s Gemini started prompting me, I didn’t choose to participate in that feature. Substack has done something similar, defaulting a setting that allows our newsletters and essays to be uploaded to a system that will train AI with it. You have to opt out in your settings.
Another problem, pointed out by tech reporter Vauhini Vara on an episode of The One You Feed podcast goes back to the very genesis of tech companies. She said the founders are often altruistic at the beginning. They have big dreams and an idea that’s going to change the world. They want to help people! Make a difference! Truly. But as soon as investors get involved, which they must, because tech products are expensive to build, the relationship to their idea immediately changes. I found this line of thinking fascinating and also very spot on.
[Side Note: Vauhini’s book, Searches: Selfhood for the Digital Age, is one of the most compelling things I’ve read about this new reality we’re in—highly recommend!]
So where does this leave us?
Can AI help me make a fun 45 minute soccer practice for my kid’s team? Sure! And that’s great when I go directly to the website, on my terms, when I want or need it. When I make the choice myself. Where I’m struggling, and I imagine you might be too, is the idea that there may come a day when we can’t actually opt out.
There’s so much I don’t know, but here’s one thing I do: Art resonates is because it was created by a person who generated their poem or painting or song after going through something. They were in it. They felt it. And they made something beautiful.
AI can’t do that. The best AI can do is fake it.
AI can write a decent line, but only because it owes all its knowledge to writers who have done the work.
When it comes to art, AI is nothing but hollow.
