What AI Can’t Do That Artists Can
By Eric Booth, January 2026[1]
AI will change everything. Or maybe it won’t. Pundits, editorialists, substackers, friends and family weigh in on how Artificial Intelligence will or won’t change this or that or every part of our lives.
And for people in the arts? In two minutes, an AI can write fifty poems about the heartbreak of being left by a lover—poems that are technically competent, with regular meter, functional metaphors, and emotionally recognizable language. Some are moving to read. It can write fifty songs in the same amount of time, songs with lyrics, chord progressions, structure, rhyme schemes, and—if your ask was smart enough—varying song structures and different genres.
Is this competition? Is this artistic Armageddon? Does the arts ecosystem need to call in the Defenders of the Galaxy to save us?
This new era is just getting started, and I am already tired of hearing about the promised Renaissance and/or looming Reckoning that AI will bring. My colleagues range from rejecting to embracing AI. I’m agnostic, and experimenting. I use Claude[2] but am committed to impersonal pronouns to describe my “relationship”…with it—Claude, I am going to resist calling you a he. It’s all a swirl. A maelstrom that may be disastrous for the arts and artists. Or not.
How can I be helpful here? This past fall, I was invited to co-design and co-lead a daylong workshop on AI for 250 business leaders attending the Inner Development Goals (IDG) Summit[3] in Stockholm. I did some exploratory learning, and, as the artist on the team, I held hard to the conviction that there must be embodied and playful learning in the day—things that artists can do that AI can’t. After several wrangling, we agreed on this as our focal question: Who are you becoming, in a world where AI is not just a tool but also a presence shaping your relationship with the world and your sense of self, purpose, and excellence?
We didn’t arrive at an answer that day, of course, but the thinking was as useful for an artist as it was for the business participants. What do people in the arts need to be exploring to find the possibilities and opportunities in the unstoppable era of AI?
In this short essay, I’ve gathered some ideas for people in the arts to consider in this time of hyper-speed AI evolution. This list isn’t prescriptive, and it certainly isn’t exhaustive or final, but it is one branch of the human tree to hold onto as the AI storm reconfigures the landscape we know.
Here are a lucky 13 things that AI cannot do that artists can. I’d like to think that AI will never be able to do these things that artists manage from their early years. Our culture doesn’t much value these particular artistic capacities…yet. But it just may be that these skills will become more valuable in our emerging culture, precisely because AI can’t provide these humanizing ways of making meaning in the world.
1. AI has no embodied knowing. AI functions only from the virtual neck up. The deeper truths and mysteries of being alive arise also from what a whole body knows. Neuroscience agrees. Indeed, neuroscientists are increasingly discovering the influence “the gut” has on many aspects of our lives; artists have deep familiarity with following the guidance of their gut. Much, if not most, creativity and creative discovery springs from the wisdom, information, and mysteries of the human body.
2. AI can’t not know. Artists are specialists at not-knowing, and we can “not know” well enough to navigate our way to new knowing. Being able to know what to do when you don’t know what to do is a valuable capacity for any complex inquiry and unprecedented challenge.
3. AI can’t understand or feel the texture of experience. AI flattens experience; artists delight in dimension. Humans require texture and feel, sometimes; in a life well lived, they often yearn for it. Humans feel familiarity; machines never will. It’s a sense so readily accessible, we build our lives around staying within it and escaping it. Artists are masters in the aesthetic medium of familiarity, playing with it in every work they create.
4. AI can’t manage the multi-colors of internal contradictions, things that normally don’t fit together or go together but do make a kind of sense in lived life and in communicating truths that go beyond the logical. Life involves internal contradictions…only all the time. Artists thrive in this mess, making new meaning with their idiosyncratic feel for new connections.
5. AI can’t start with a shard of something (a sensation, a memory, a sliver of sound or language, a hunch, a remembered taste of a childhood cookie) and follow an intuitive pathway from there toward something new—something that might or might not be significant to others but is a destination only a living individual could find.
6. AI can’t get obsessed by something and keep working on it over time, from different angles. Artists keep going deeper, working on it, playing with it, because it derives from who they are and the mark they want to make on the world.
7. AI can’t delight in and make use of odd synaptic connections. AI can’t grasp the flash of green that reminds you of a dress your mother wore to a holiday party where you were proud of the way people listened to her when she told stories that included you. And AI can’t then tell that story in a way that makes others remember their moms at social gatherings and then love their long-gone mothers in a deeper way.
8. AI doesn’t know how to break rules effectively. Artists can feel when and how to break an expectation or a norm, to get a better, fresher, more impactful outcome.
9. AI doesn’t know how to think through the medium it’s working in. Artists know how to let the clay tell them what to do, or how to let the sound of a particular guitar open up something new.
10. Both AI and artists work well within constraints. But AI uses constraints as optimization parameters to get a job done; artists can embrace them as enabling constraints, as creative partners, that take them to ideas they wouldn’t have had otherwise, to higher quality and to greater meaning.
11. AI can’t create out of the personal experience of grief, or of love, and it has no feel for legacy. AI can cull from and generate descriptions of experiences of our most profound emotions, but it can’t yearn its way to the exact expression that helps us hold the ineffable. AI can’t hope.
12. AI can’t think sloppily. AI can’t make a mess, intentionally, to discover original ideas out of a jumble or out of mistakes. Artists know: “Not all who wander are lost.”[4]
13. AI can’t reach for the tool of playfulness when trying to address a difficult problem. I asked Claude about its ability to be playful. The answer: “I can simulate playfulness effectively when that’s the job. But I can’t spontaneously become playful as a strategy when I'm stuck, or as a response to absurdity, or as a relationship with you. It’s like the difference between an actor playing someone who’s laughing versus someone actually laughing at a joke. I can do the former convincingly. I cannot do the latter.”
That thirteen is just a start. I’m not even counting the thousand things that performing artists can do that AI can’t. How the good storyteller can ignite an audience at a Stockholm conference in a way that hours of the AI’s information can’t. How an end-of-day conference summary poem by an AI, in the style of Alan Ginsberg, can be impressively clever and complete, but not make us lean into our seated neighbor, touch her hand, and start talking about what matters to us, the way the human poet’s did.
I asked Claude to offer a closing thought for this essay. I prompted it with: What is the most important thing I can tell people in the arts about their relationship to and future with AI? The AI answered:
AI is a collaborator that executes, not a consciousness that creates meaning.
Artists make choices that matter because of who they are—their lived experience, their specific obsessions, their accumulated taste, their point of view. AI has none of that. It can generate technically impressive work, but it cannot decide what needs to be said or why.
The artist's irreplaceable role is intentionality, knowing what you're trying to say and why it matters. AI can help you say it in a thousand different ways, but it can't tell you which one is true to your vision. It can't feel the difference between ‘close enough’ and ‘exactly right.’
Your competitive advantage isn't manual dexterity or even technical knowledge anymore. It's your ability to recognize what's worth making and why.
Claude is smart, isn’t it? Its ideas are helpful, aren’t they? I’m excited to have Claude as a helper. In my lifetime in the arts, this is the first time I’ve ever been able to afford a creative assistant; and this first one is damn good, within the stated limitations.
Here is a question for us in the arts to wrangle with, using our 13 strengths to find a workable way forward, as we always have. In a world in which AI does soulless things that resemble what we have done for 30,000 years, but faster and cheaper than we can, how do we enjoy doubling down on our artistic excellences, to discover new things that matter in the world we yearn to create?
[1] Written while in residence at The Hermitage Artist Retreat, with thanks for their support.
[2] Claude is an AI program of the company Anthropic
[3] Inner Development Goals/IDGs (https://innerdevelopmentgoals.org/). ITAC/International Teaching Artist Collaborative, an organization I co-founded in 2012, serves as the Artistic Director of IDG Summits. My co-designers and co-leaders for the AI track were the founders of Sutra (Lorenz and Natasha Sell) and Pontus Holmgren of Gro
[4] J.R.R. Tolkien, from The Lord of the Rings