What AI changed about how we encounter art
It's been a few years now since the advent of AI‑generated art and music. Our world, for better or worse, is changed now.
AI art has reshaped our relationship with art by breaking a long-held assumption.
That what we encounter was made intentionally by another person.
When images can be generated endlessly, without effort or care, intention becomes unreadable.
Art shifts from an invitation to a human experience to a mechanism for extracting attention.
I have been following the progress of OpenAI since 2018 with a lot of anticipation. I was fascinated by machine learning and the breakneck pace of advancement. I was amazed at how many applications they had, and when DALLE-2 dropped I was granted early access.
An expanded invitation to create. Practically overnight any human was granted the ability to create an image beyond their physical skill. All it took was the ability to describe a thought. There is something that I still find comforting in that, but I must admit that the original novelty has worn off.
I saw all kinds of takes at the time, where people were understandably threatened — threatened for their livelihood. Many offended by the idea that a machine could create art at all.
We argued about authorship, not attention. I have never been employed as an artist, but I make art. I understood that affront to our egos, but was reminded of an article that I read many years prior in which a computer scientist generated compositions in the style of Bach and had them performed by a human orchestra. Audiences described the music as 'beautiful' and 'inspired' before they believed Bach had composed it but 'lifeless' and 'plain' when they knew it was created by a computer.
Surely if the difference was knowing that a computer produced something, then the real issue was our egos thinking so highly of our human capability and creativity.
Play, intention, and early joy. I enjoyed using AI to create something new and beautiful. I even hosted a cat-themed party when MidJourney first got good. We invited friends to come and generate images of cats of all kinds. We printed them out and put them on the wall where they still hang today.
I had a lot of thoughts about how we could monetize the new capability to not be in fundamental contradiction to the wellbeing of artists. Perhaps we could devise a system where to train a model where artists could elect to have some small sum deposited in a crypto-wallet on each generation, or a way to legislate opting out of the data collection altogether that had teeth. still think those are good ideas.
But all in all I would say that I was excited by the technology, and intrigued by its capabilities. I wanted to see all the new and powerful things that AI could unlock for us, especially new forms of art not yet devised.
In his 1972 documentary, Ways Of Seeing, Johnathan Berger makes the case that something fundamental changed about art with the invention of the camera. Before the camera, Berger argues, art was always by design. If you wanted to see a fresco in the 1600s you would have to enter a church or a museum. Every person that wanted to see that painting would also have to walk the same staircases, the same hallways. They would see it in the same frame, with the same background and the same lighting. Art was sacred.
If you saw a painting you would know that every detail of its presentation was controlled and understood by the artist. If your friend told you about some wonderful artwork they had beheld, you would only have their description to imagine it by.
But with the camera and the modern printing press, classical artists lost their setting. They lost their control of their art and how it was viewed. Suddenly, these once-sacred works of classical art were used as a prop to advertise in magazines and newspapers.
What Berger was really describing was not the loss of intention, but a shift in who controlled it.
When images could be detached from their original setting and reproduced freely, artists lost authority over how their work was encountered.
But intention still came first. Someone still chose to reproduce the image, to place it, to use it. Even misuse was deliberate.
What changed was not whether intention existed, but where it lived.
I posit that AI capabilities have similarly changed something about art and how it is consumed. Where the camera redistributed control over intentionality, AI dissolves the assumption that intention preceded what we encounter at all.
The candy rock moment. I find myself, like most of our species, sometimes scrolling my life away on social media with no intention behind my own actions. And while I am scrolling, I come across a comic. I see an apparent narrative arc, but the words make no sense. The algorithm has given my dopamine-addicted brain a morsel to chew on, and it seeks to understand.
But as I look it becomes clear: the comic is AI, and not just in structure, but in total. Every comment comes from a fresh account only minutes old and the image itself is just nonsense. And what's more is that I stopped and tried to parse the image. I sent the signal to send me more things like this. Social media wants me doing this to maximize ad revenue.
What unsettled me most was realizing that no one had decided this was worth making before I saw it. My attention wasn’t invited into an experience—it was being probed to see whether meaning could be extracted after the fact.
It was as though someone promised to give me a candy but instead handed me a sugar-covered rock. Realizing the way in which I was being manipulated caused me to regroup and put down my phone. It really got me thinking.
Before AI art, all art we saw existed because some human had decided to make it. No matter how small, it cost the artist a sliver of their lifespan.
But now, a marketer with no artistic skill or intention, can create infinite art and put it in front of your eyes for you to react to. And they will.
This phenomenon will not stop at digital and printed media. Today there are already products on the market that can create physical art with minimal human input. As time goes on the, line between intentional art and unintentional art will blur.
Do not get me wrong: the endeavor to automate art is an art form of its own. I could imagine truly intentional uses for such a machine, and new capabilities realized from it. But this new form of art has a new ability: it can pass itself off as being intentional when there was no intent.
I hadn't given due credit to art before, but its very nature bestowed a Trust (Capital T) – Trust enough that the artist thought it was worthwhile to create. And now that Trust is being hijacked.
We can no longer just go through our lives believing that what we are beholding has a true human intention behind it.
The good news is we still have choices.
The reality is that we can choose to not make use of this tool in our creative processes when it is not necessary. Just as the camera did not remove our ability to designate camera-free spaces and generate art that could not be captured in photographs, AI does not remove our ability to use our human intention to direct artistic output.
If you are similarly afflicted, and desire trust in art as a window for human connection; If you desire a world where art once again communicates the thoughts of people – where beholding it instantly transports us across time and space, back to an actual moment that another human experienced. Then we have a few levers we can pull.
I, personally, don't want "No AI ever" and I don't feel alone in my admiration for some AI capabilities. There are places where AI makes sense to use and makes life better, including art generation. I use AI to make programming more accessible and to test design decisions, as is the case for much of this site. It also helped immensely when I needed to cross language barriers while traveling.
However, there are places where AI art is not needed or wanted. When a marketer uses an AI image to avoid paying a small sum to someone that took stock photos that already exist and on which the AI model is trained, that sucks. When propagandists make doorcam videos of crimes that never happened and get people pumping cortisol through their veins over something that never happened, that sucks.
Restoring signals of intention. I believe we can build support around causes like this and legislate specific use cases. Progress can be made in our society to regulate the ways people are using AI, and disclose when it has been used. To do this we would have to have some common understanding of where the line is for using AI art and aligning ourselves with that ideal.
Refusing extraction as a default. If we don't want this, then it is up to us to resist doing these things in our personal or professional lives. This is why I have made the following commitment.
Here is a version of that commitment:
I hope that you see that I have already broken this commitment by making this article. This was a deliberate choice. We're not going to be perfect and perfection isn't the goal. We want to establish culture and shared values, not be cops to one another. Our world is crazy. But by divesting ourselves from these malicious use cases, we pave the way to legislate.
If you see the value in what I'm saying then consider making the commitment. There's no club, but if you are serious about it then consider getting out a pen and paper and writing down a few of your commitments that you can stand by and sign them. Be reminded of what you believe.
Art used to arrive with an unspoken promise: that someone, somewhere, decided it was worth their time to make. That promise is no longer guaranteed.
AI didn’t remove intention—but it made it optional. And when intention becomes optional, trust does too. What once invited us into another human experience can now exist solely to extract our attention, without care for whether it was worth making at all.
We still have agency. We can choose restraint over convenience, disclosure over ambiguity, and invitation over extraction. The future of art won’t be decided by the tools we invent, but by the norms we’re willing to uphold while using them.