I read an article that said AI is beating the average human in creativity tests. You heard that right, it has been shown that advanced AI models can now outperform average humans in standardized creativity tests.
Sometimes, when playing a character in a play or a film, I have a moment, if all is working, if I have done my homework and I am truly allowing the character to breathe, when I suddenly leave my body, and I am quite literally, in the back of the house watching myself perform. It’s not frightening; it doesn’t feel out of control, and when it happens, I have no idea how or why. This happened to me recently while playing Scrooge. In the end, when Scrooge has his epiphany and realizes he has changed, and he has missed so much, I was suddenly in the back of the house watching myself on stage. It was crazy.
I share this story because it was what I thought of when I read an article that said AI is beating the average human in creativity tests. You heard that right, it has been shown that advanced AI models can now outperform average humans in standardized creativity tests. That sounds terrible to me. The article did go on to say that elite human creators still have a significant advantage.
How do they define elite? What do these tests really measure? And what results are we actually looking for? Those are a few questions I’d like to ask, as a creative, as someone who is worried they will be replaced by an algorithm. But I also wonder why I’m asking these questions. What will I do with the information? What will anyone do with the information? And more importantly, what CAN we do with this information?
WHAT THOSE CREATIVITY TESTS ACTUALLY MEASURE
When researchers say AI is outperforming humans in creativity tests, it helps to understand what those tests are actually asking for. They aren’t asking whether an idea resonates, whether it changes anyone, or whether it costs the person creating it something internal. They’re asking for output. Usually a lot of it. Under a defined constraint. In a short amount of time.
Most standardized creativity tests reward volume, variation, and statistical rarity. Originality, in this context, doesn’t mean meaningful. It means uncommon compared to other responses. That’s an environment AI thrives in.
AI has absorbed more examples than any human ever could. So when you ask it for alternatives, it doesn’t hesitate or second-guess. It isn’t attached to what it produces. It just keeps going. If you’re scoring speed and variety, that looks like creativity. But creative work in the real world rarely works that way.
Most of the time, creativity isn’t about producing as many options as possible. It’s about deciding which option matters, sensing when something is technically impressive but emotionally empty, and knowing when to stop generating and start committing.
Those decisions don’t show up on a standardized test. They can’t be measured without context, consequence, or time. And they don’t capture what happens when a piece of work begins to change the person making it.
That’s something I have never heard an AI specialist or advocate ever mention: the conversation between the art and the artist. Because that exists. And it exists in every painter’s studio, dance studio, rehearsal hall, garage, or field where art is being created. The artist has an impulse. The impulse needs to be served. And as it’s served, it becomes clearer. What they want to say. What they’re actually feeling. And, most importantly, the artist listens to and watches their work, hearing what it’s telling them. From my point of view, if I am doing a comedy, we use the previews to feel the audience; do they laugh here, are they getting the story here, are these plot points clear? We listen to the art we create. It helps us find direction, specificity, and message. We adjust based on what we feel.
So yes, AI can win at creativity tests. But those tests are measuring a narrow slice of the process. They measure exploration, not judgment. Possibility, not meaning.
Which matters, especially before we decide what it means to “lose.”
WHAT RESEARCHERS MEAN WHEN THEY SAY “ELITE CREATIVE”
When the article says elite creatives still hold an advantage over AI, it’s tempting to assume it’s talking about fame, money, or some visible level of success. The kind you can point to. The kind that makes the word elite feel earned. That’s not usually what researchers mean.
In studies, elite creatives tend to be defined less by recognition and more by consistency. Not how many ideas they generate, but how reliably they can produce work that holds up under pressure, inside real constraints, with real consequences.
They’re the people who can enter a messy problem and make it clearer. Who can sense which idea is worth pursuing before there’s proof. Who know when something is almost right and when it’s fundamentally wrong, even if they can’t immediately explain why. This isn’t about inspiration. Elite creatives don’t spend most of their time generating ideas. They spend it choosing, shaping, discarding, and sitting with discomfort longer than most people are willing to. They let things stay unfinished until they reveal themselves.
That ability doesn’t show up on a test designed to reward speed or novelty. It shows up over time, across projects, across failures, in how often someone’s instincts turn out to be right. Which is why elite creativity is so hard to measure and so easy to misunderstand.
From the outside, it can look slow, overly particular, even resistant. But what’s actually happening is filtration, a narrowing, fewer ideas held to a higher standard.
AI is excellent at producing options. Elite creatives are practiced at deciding which option is worth living with.
On Sitting With Discomfort
Discomfort and fear are two emotions that AI can technically replicate, but it cannot feel. Again, I’ll use the analogy of doing a comedy on stage. Say it’s all going well. Perfect cast, solid timing, clear direction, moments to pause, moments to drop the laugh. Everything feels locked in. Every actor feels their legs under them solidly. The director has fine-tuned the show so that it’s like a finely tuned watch. However, before the audience arrives, the actors feel a bit of panic.
When you’re doing a play, the audience is always the third character. The play is not fully complete until an audience shows up and shares their energy, laughter, sorrow, whatever and then the actors take that information, that data, and they adjust the play. This is especially true when doing a comedy. You learn when to hold for laughs, you discover, when an audience arrives, things that are funny that you never thought would be. The third actor, the audience, gives the performers so much data to use, not use, feather in, and complete the show. Any actor will tell you that after rehearsing a comedy in an empty room, you suddenly feel worried, will this bit work, will this joke land? You know in your bones you’ve done the work, and you understand the rhythm, the beats, the timing, but after throwing it all into an empty room, you start to doubt. You start to question. Those feelings drive the work forward, and part of that work is feeling deeply uncomfortable because that one piece is still missing. So, you’re forced to sit in that discomfort and give it its full breath and trust that you’ll come out the other side and you’ll survive.
But the discomfort is part of the process and a vital one at that. You learn over time and experience that either this discomfort might be telling you something good, or it’s just preshow discomfort, and you breathe through it. Either outcome is important to the crafting of the art. You need to feel whether you’re being honest and not just giving a vague simulacrum of what an emotion is. That’s what AI does. Shows you what the emotion is, but cannot tell you what it feels like. Creatives do that. Not just the elites, but all creatives do that.
DOES ELITE CREATIVITY REQUIRE SUCCESS, MONEY, OR FAME?
This is usually the quiet question underneath all of this. If elite creatives still matter, does that mean only the ones who made it matter? The visible ones. The rewarded ones. The people with proof.
It’s an understandable assumption. We’re trained to look for outcomes as validation. If someone is successful, we assume their creativity must be elite. If they’re not, we assume it wasn’t. But that logic runs backward.
Success, money, and fame are not requirements for elite creativity. They’re side effects that sometimes show up when elite judgment survives long enough in public. Sometimes they never show up at all. Elite creativity is not a status you reach. It’s a way of working that compounds.
It shows up in how often someone can return to the work without shortcuts. In how willing they are to be wrong early, so they don’t ship something dishonest later. In how consistently they can tell the difference between something that’s impressive and something that’s true.
Many elite creatives never become famous because their work lives inside systems, teams, and processes. Inside rehearsals and drafts and rooms where no one ever applauds. Their value isn’t broadcast. It’s relied on. Money doesn’t teach that. Recognition doesn’t either. Time does. Failure does. Repetition does.
Elite creatives tend to have spent years making things that didn’t work, sitting with that discomfort we just described, learning how to tell whether the feeling in their gut is fear or a signal. That’s not a talent. It’s a practiced sensitivity. Which is why creativity tests often miss them.
Tests reward speed. Elite creatives often slow things down. Tests reward novelty. Elite creatives are suspicious of novelty that hasn’t earned its place. Tests reward production. Elite creatives know when not to produce yet. None of that looks impressive in a benchmark. But it’s the reason certain people become anchors in creative processes, whether anyone knows their name or not.
And it’s the reason replacing creativity with output efficiency doesn’t just change who wins. It changes what the work becomes.
WHERE AI STOPS, AND WHY THAT LIMIT MATTERS
At this point, it’s tempting to say that AI just needs more time. More training. Better models. That eventually it will cross whatever line still separates it from human creativity. But that assumes the gap is technical. It isn’t.
AI doesn’t lack information. It lacks consequence. It doesn’t feel the weight of choosing wrong. It doesn’t carry the residue of past failures into the next attempt. It doesn’t hesitate because hesitation costs nothing.
When AI generates, nothing is at stake. There’s no internal reckoning. No moment where a decision lands in the body and has to be lived with. It can revise endlessly without embarrassment, without regret, without relief. That absence changes everything.
Creative work isn’t just about producing something new. It’s about navigating uncertainty while being affected by it. It’s about making decisions that feel risky because they are. It’s about committing to something before you know if it will work, and then standing inside the result. AI can simulate uncertainty. It can describe risk. But it doesn’t experience either one.
Which means it can’t experience relief either. Or transformation. Or the strange quiet that sometimes arrives when something finally clicks and you realize you’re not the same person you were when you started. That’s not a training gap. That’s an existential one.
And it’s why moments like the one I described earlier don’t register on any benchmark. There’s no metric for leaving your body and watching yourself perform. No score for the internal shift that happens when the work changes you as much as you shape it. Those moments aren’t outputs. They’re costs.
They’re the price of caring enough to be altered by what you’re making. And that price is something AI will never pay, not because it isn’t advanced enough, but because there’s nothing there to be altered.
Which raises a question that has less to do with technology and more to do with us.
When Creatives Talk Cost
“What does this moment cost you?” asked my Grad School acting teacher as I worked through a scene in Clifford Odets’ Golden Boy. What did it cost me? I was a bit confused because I equated cost with money or property. I would soon learn that cost was an emotional thing. What did it cost me, as the character, to admit something so personal to this woman I just met? I learned that giving yourself over causes you to worry about being accepted, understood, mocked, humiliated, or rejected. Now, all those things add to the emotional cost of the moment. The higher the cost, the deeper the emotional response.
What a moment costs changes as the actor adjusts to the scene. As the rehearsal progresses, the cost becomes greater, the risk becomes greater, and the fear deepens, and that’s where the friction comes in, the struggle, the real emotions work. That’s the conversation between the artist (actor) and the art.
Humans have these conversations all the time: “I should work out, but I just want to go home and veg on the couch.” The human emotions on display here are guilt, worry, and the need for self-care. All of these have an emotional cost. Then you add in the voice that says, “You’ll feel better once you’ve worked out.” What is the cost of ignoring that voice?
When something costs us emotionally, we view it differently. A high emotional cost demands a high level of commitment. For an artist, everything must have a high emotional cost, or why bother painting it, dancing it, singing it, acting it? Art isn’t about the mundane; even if the painting looks mundane, or the scene in the film is just a guy working in the garden, art shows us the heightened moments in life. Those heightened moments are often hidden or repressed in real life, but art allows them to be seen and heard. At a cost. And that feeling of cost cannot be replicated by AI because it has to be felt.
When we set out to create art, write a play, act, sing, dance, we set out to show the heightened states of the human experience, the universals, and how those make us feel and what we do with those feelings. Again, this is the conversation between artist and art, a conversation that AI models will never have.
DO WE KEEP FEEDING WHAT AI CAN’T FEEL, OR REDEFINE CREATIVITY?
This is the part that makes people uncomfortable, and it should. Hell, it makes me feel uncomfortable, and I’m writing this.
If AI can generate endlessly without fear, without hesitation, without emotional cost, then the question isn’t whether it’s creative enough. The question is whether we still value the parts of creativity that slow things down. Because those parts are inefficient.
Feeling costs time. Sitting with uncertainty costs momentum. Allowing work to change you costs control. None of that scales well. None of it fits neatly into benchmarks or dashboards. So there’s a quiet temptation to redefine creativity as something cleaner, faster, or closer to output. If the result looks good and gets attention, does it matter how it was made? Or what it costs to make it? That’s not a future problem. It’s already happening.
We reward speed. We reward volume. We reward cleverness that lands quickly and moves on. And the more we do that, the more we train ourselves out of the very skills we claim to be protecting. Judgment, taste, and patience. The ability to stay with a feeling long enough for it to tell you something true. AI doesn’t erode those skills directly. We do. By choosing convenience over craft. By choosing abundance over meaning. By deciding that if something can be generated, it doesn’t need to be lived.
If creativity becomes defined by how much can be produced, AI wins easily. If it’s defined by how deeply something is felt, understood, and shaped in response, then the work stays human. But only if we keep practicing those capacities. That’s the part no model can automate.
The real risk isn’t replacement. It’s neglect. Letting the muscles that make creativity human weaken because something else can carry the load faster.
Which brings us back to the moment on stage.
Let’s Talk Fear
Another lovely little human nugget that AI cannot replicate because it cannot experience it is fear. Fear is good. When we deal with our lizard brain, the amygdala, we have fight or flight, and those instincts keep us alive. We need fear to keep us from licking the hot iron or warming our faces on a lit gas stove. But we also need to know fear to overcome fear.
When we care about art or performance, we fear it won't be accepted, will be harshly reviewed, or won't be understood. As creatives, once we express and own that fear, we can usually overcome it through experience. I know that on opening night, I will be full of fears, but once I step on stage, the fears go away. And I know this happens because I care, and I want to tell a good story for the audience.
Creatives connect fear with care and importance. If we weren’t worried or afraid, it usually means we don’t care enough about the piece we’re engaged in creating or have created. Fear is good, and you can only describe it, write about, and convey it to an audience if you’ve experienced it. Again, AI cannot experience anything.
BACK TO THE MOMENT ON STAGE
When I think back to that moment playing Scrooge, the strange clarity of watching myself from the back of the house, what stays with me isn’t the performance. It’s the feeling that something had moved. Something internal had shifted. The work wasn’t just working. It was changing me while I was inside it.
That moment didn’t arrive because I found the right interpretation or made a clever choice. It arrived because of everything that came before it. The preparation, fear, and the emotional cost. The willingness to be seen and possibly misunderstood. The care I had for that moment. AI can describe that moment, analyze it, and even generate language that sounds like it understands it. But it will never stand in it. It can never live it.
It will never feel the quiet panic before the audience arrives. It will never feel fear dissolve into focus because the story matters. It will never leave its body because there is no body to leave. There is no internal state to be altered by the work itself. That difference matters more than any creativity test score.
Because creativity, at its most human, is not just about making something new. It’s about allowing the process to cost you something and trusting that cost to shape the work. It’s about feedback loops that run through nerves, doubt, care, and commitment. It’s about moments that can’t be optimized because they can’t be predicted.
And if we lose sight of that, not because AI took it from us but because we stopped valuing it, then creativity doesn’t disappear. It just becomes thinner. Safer. Easier to generate and harder to feel.
That’s the real risk hiding underneath the headlines.
THE TAKEAWAY
AI isn’t ending creativity. It’s ending the scarcity of ideas.
What’s becoming rare instead is judgment, taste, the willingness to sit with discomfort, and the ability to let work change you before you ask it to change anyone else.
At ThoughtLab, we don’t see creativity as output. We see it as a practice. One that requires attention, consequence, and care. Especially now. Especially when speed and scale are so tempting.
The question isn’t whether AI can be creative.
It’s whether we’re willing to keep feeding the parts of creativity that only humans can feel.
