A painter friend of mine, who's been doing gallery shows for almost fifteen years now, told me one thing that stuck with me from last winter that I wouldn't expect it to still be resonating with me so much. The scariest part of any project was never finishing it," she said. It was fifteen minutes before she picked up a brush, before the canvas was blank, before all the choices were equally plausible and inexplicably equally problematic at the same time. After that, she has begun to use AI to create quick sketches to overcome that crushing sense of paralysis before she can get to her real work and nothing more.
It's not quite the appeal one might imagine when saying "artists and AI.If you've seen "artists and AI" in the same sentence, that's not the appeal you'd imagine. All the chatter on the web is about robots replacing painters, novelists getting left behind, some dark day in the future where no one will ever hold a pen again. What really people are doing in a lot of studios, and a lot of writing desks, right now, looks nothing like that. It seems like what my friend does. The tool to remove the fog a little so that the other, human part of the work can get going.
The Fear Was Never Really About Creativity
So, let's get back to the origins of all the anxiety. When AI began to generate pictures and text that seemed to be okay, a lot of artists panicked. And here is why: the fear was not without reason. No one learns a craft to work for a decade and then finds a machine can do it in 4 seconds.
But when the initial panic subsided a bit, something interesting happened. Many working artists and writers began to observe that any work generated by AI without any human control, direction, or guidance sounds a bit flat. Sure, it's technically fine, if not even impressive at first sight. Yet without any of the sensation that the piece originated in a place that actually exists. A specific memory. An unusual fascination which someone developed from childhood. Artist's never-let-go grudge. There's no such thing as that inside of AI. It's taking patterns it has seen a million times, but it's not from an actual life that was actually lived.
I can't say that changed the whole conversation, right? A lot of creative folks began asking a much better question when they first started using this thing: what if I just use it like any other tool and it will take care of some of the chores so I can get to the part I can do?
Where AI Actually Helps, According to People Who Actually Use It
I've discussed it with plenty of writers, and a few visual artists as well, and I am beginning to see a trend in how they actually use these tools. Which is frequently much different from what people think they use.
AI-powered writing tools for writers are much more likely to be present at the beginning of the process, not the end. No serious writer wants to go into AI to write an entire novel and have their name on the cover. In reality, what happens is you are stuck on a plot point, and you push your imagination to come up with five wildly different directions for the story to go, just to make something happen in your head. Or you're in the situation where you just have to ask yourself why this guy does this thing? And you have to explain it to him and he provides you with some decent questions after that. That's still the writer's prose; that's the sentences that make it into the final draft. Every time. The tool is more of a sparring partner than a ghost writer, if you will.
The same is said by visual artists. Many of them use the images created by AI as raw material to respond to and, much of the time, get aggravated about, rather than the actual product. Sometimes, an artist will come up with a hundred sketchy compositions, despises eleven and then finds one weird little feature in the twelfth one that inspires them to start painting something that isn't remotely similar to the AI's initial work. That painting was not painted by the AI. Just passed on something to argue about.
Academic and Long-Form Writers Are Finding Something Pretty Similar
This isn't limited to poetry and painting, either, which surprised me a little when I first started noticing it. A lot of the same dynamic shows up in academic content creation, which sounds like a strange place to bring up creativity until you actually sit and watch someone write a dissertation for six months straight.
Academic writing has its own weird creative problem, if you think about it: how do you take a genuinely original idea and force it into a structure rigid enough that some committee will actually sign off on it? That's a real creative tension, and it's exhausting, genuinely exhausting, in a way people outside academia rarely appreciate. Researchers using AI text generation tools for early drafts describe basically the same thing artists describe to me. It's not that the tool writes the argument for them. It's that having something to push back against, some rough draft you get to disagree with, makes the actual thinking happen faster than staring at a blank page ever managed to.
Same goes for tools that work as an
academic paper generator. The decent ones don't hand someone a finished, polished paper they can just submit and call it a day. What they actually do is take a pile of scattered research notes and shape them into something with enough of a skeleton that a writer can start disagreeing with it, moving pieces around, sharpening whatever argument was buried underneath the mess the whole time. The creative labor, the part that actually matters here, stays entirely human. The tool just gets you to the argument faster than you'd get there alone.
The Collaboration Only Works One Way
The most important one and that's easy to overlook if you're not paying close attention. Each of the examples listed above works only because a human is driving the whole show and ultimately determines what's good.
There are 100 renditions of anything that AI can regurgitate. It doesn't, however, have a clue which one is of interest. It does not realize that there is a great deal more emotion in a sentence that is a little clumsy than in a sentence that is highly polished; that an ugly, awkward brush stroke can tell you much more than a clean one ever can. This is a judgment; this is the actual taste that makes something worth making in the first place, that is still completely in the fingers of the one who holds the pen or the brush. The only thing that has changed is how quickly someone can go from, "I have no idea what to do, where do I begin?," to, "okay, now I have something to push against."
A few things keep coming up, over and over, when creative people try to explain why this actually works for them:
● It kills the blank page problem, which has probably always been the single biggest creativity killer out there
● It gives them more raw material to react against, way faster than grinding it out solo ever allowed
● It doesn't replace their voice; it mostly just clears space for their voice to actually show up
● It speeds up the boring, structural parts of a project, leaving more energy left over for the parts that genuinely need real creative thought
Where This Leaves Artists Going Forward
Nobody serious thinks AI is going to replace the actual creative act, the part where a person decides something matters enough to fight for and drags it into existence anyway. What's shifting is the boring, grinding front half of the process, the part that used to eat up hours before the real work even had a chance to begin.
My painter friend still spends most of her time doing exactly what she's always done. Standing in front of a canvas, making decisions nobody else could possibly make for her, using the specific, weird judgment it took her fifteen years to build up in the first place. The AI just helps her get past that awful first fifteen minutes of staring at nothing. Everything after that is still entirely, completely hers.
Final Thought
The artists and writers actually embracing this stuff aren't looking for some shortcut around the hard part. They're the ones who already figured out where the genuinely hard part actually lives, and it was never in generating a rough first pass. It's in the taste. The judgment. The stubbornness to keep working on something until it finally says what you actually meant to say. AI can hand you a hundred starting points, no problem. It still can't tell you which one's worth finishing. And honestly, that's the whole job right there.
