0
0
Somewhere in the last three years, a new reflex showed up. You see a picture, a song, a short film, a paragraph that lands — and before you decide whether you like it, you check something else first. You ask whether a person actually made it.
It is a strange question to have to ask. It was never a question before. And it is now the human-made vs AI question sitting in front of almost everything we look at, quietly deciding what a thing is worth before we have really looked at it.
Here is what makes it interesting. Researchers have been testing that reflex for a few years now. What they keep finding is not quite the story most of us assume.
What Happens When You Change Only the Label
The cleanest way to test this is to change nothing about the work and change only the sentence underneath it.
That is exactly what one team did. Kobe Millet, Florian Buehler, Guanzhong Du and Michail Kokkoris ran four experiments with 1,708 participants, published in Computers in Human Behavior under the title Defending humankind: Anthropocentric bias in the appreciation of AI art. Same artwork. Same pixels. The only variable was whether people were told a human or a machine made it.
When people were told a machine made it, they liked it less. They judged it less creative. And because they judged it less creative, they felt less awe looking at it.
The effect was strongest in one particular group: people who already believe creativity is something only humans can do. Which is worth sitting with, because it means the judgment was arriving before the looking.
Human-Made vs AI: The Part Nobody Expected
Then the numbers got sharper.
Researchers at Columbia Business School put together 28 images — half of them lesser-known works by artists including Paul Gauguin and Andy Warhol, half of them generated to resemble those same styles. Participants valued the art labelled as AI-generated about 62 percent lower than the art labelled human-made.
So far, so predictable. But look at what those same participants were willing to admit.
They conceded the AI-generated work showed the same level of skill. They conceded it showed the same nuance. They did not claim it was clumsy or crude or obviously fake. They looked at the craft and called it equal.
They only marked it down in three places: creativity, the amount of labour they assumed went into it, and money.
Read that again slowly. The downgrade was never about the object. Nobody was looking at the work and seeing something worse. They were looking at the work, seeing something equal, and then pricing it as though it were not.
The Label Only Works in One Direction
Here is the finding that quietly undoes the whole comfortable version of this story.
In June 2026, Han Sol Lim, Bong Gyou Lee, Yoon Hi Sung and Chang Won Jung published a study of 618 participants in Frontiers in Psychology testing how provenance labels — “human-made,” “AI-generated,” or nothing at all — change how much effort people believe went into a piece of content.
The “AI-generated” label did what you would expect. It significantly reduced perceived effort.
The “human-made” label did nothing. It produced no meaningful lift compared with content that carried no label at all.
We have found a way to mark machine work down. We have not, so far, found a way to mark human work up.
The researchers were careful about their own result, which is a good sign in anyone. They noted the finding may reflect the current, transitional stage of this technology, and could shift if machine-made work ever becomes the default rather than the exception. That is an honest caveat and it is worth keeping. But as things stand, “made by a human” is not a promotion. It is only the absence of a demotion.
Human Work Only Won When the Machine Was in the Room
One more detail from the Columbia research, and it is the one that has stayed with me longest.
Human-made art was rated as more valuable when it was displayed next to AI-generated art. But when the same human-made art was displayed next to other human-made art, that lift disappeared.
The value was not sitting in the work. It was sitting in the contrast.
And that is a fragile way to hold worth. Something that only looks valuable when there is a lesser thing standing beside it is not really being valued at all — it is being ranked. Rankings move every time the room changes. Bring in one better painting, one cheaper competitor, one person who does your job slightly faster, and the number resets.
Which is strange, because people have been making things for a very long time without any of this. Somebody in an Ice Age cave, cold and hungry and short on daylight, still stopped to carve a bird the size of a thumbnail out of ivory, roughly forty thousand years ago. There was no gallery. There was no label. There was nothing to compare it to. Whatever made that worth doing, it was not the ranking.
What We Are Actually Measuring
Put all three studies next to each other and the same shape keeps appearing.
The image does not change. The brushwork does not change. The skill is granted. What changes is the answer to a completely different question — was there an intention behind this?
That is what people are reaching for when they ask whether a person made it. Not craft, because they already handed the machine the craft. They are reaching for the fact of someone having meant it.
Which is an old instinct, older than any of this. The oldest account we have of where human beings came from does not open by measuring what a person can produce, or how skilled they are, or how they compare to whoever is standing next to them. It opens with the claim that a person was made deliberately, by God, on purpose — and that the worth was settled right there at the making, before that person had produced a single thing worth looking at. There is an old word for that idea, and it means, roughly, a thing someone made on purpose.
It is a completely different accounting system than the one we are currently running. Worth conferred at the making, rather than assigned afterward by comparison.
The Test We Run on Ourselves
None of this means the reflex is wrong. It is reasonable to want to know where a thing came from. Knowing who made something, and what it cost them, is real information and it belongs in the decision.
But it is worth noticing what that reflex costs when we turn it inward — because most of us do. We run the same audit on our own lives. We check what we produced this year. We check it against whoever is standing next to us. And we let the ranking decide what we are worth.
The research says that test is unreliable even on paintings. It says the exact same object gets two different valuations depending on nothing but the story told about it. It says the number moves when the room changes, not when the thing changes.
If you have ever found yourself doing that arithmetic at two in the morning, the free What Is My Purpose? assessment is a gentler place to start than the comparison — it asks what you were built for rather than how you rank.
So the next time you catch yourself checking whether a person made something before deciding whether it matters, notice what you have already assumed. You have assumed that being made, on purpose, by someone who meant it, changes what a thing is worth.
Then notice that you were made too.
A Question Worth Sitting With
Here is the one I keep going back and forth on. If you genuinely could not tell the difference — the song moved you, the picture stopped you, the story stayed with you for a week — should it matter who or what made it?
Some people say the experience is the entire point, and anything else is snobbery. Others say a thing made without intention is a different kind of thing no matter how good it is, and we are right to feel that. Where do you land? Leave a comment and tell me which side you are on, and what made you land there.
If This Was Worth Reading, Pass It On
- People rated the exact same artwork 62% lower once they were told a machine made it — while admitting it showed the same skill and the same nuance. We are not judging the work. We are judging whether somebody meant it. https://bgodinspired.com/index.php/entertainment/human-made-vs-ai-why-the-maker-changes-the-value/
- Turns out “made by a human” is not actually a selling point. In a 2026 study of 618 people, the “AI-generated” label dropped how much effort people thought went into a piece — but the “human-made” label added nothing at all compared to no label. We know how to mark the machine down. We have not worked out how to mark the person up. https://bgodinspired.com/index.php/entertainment/human-made-vs-ai-why-the-maker-changes-the-value/
- The finding that got me: human-made art only gained value when it was hanging next to AI art. Next to other human art, the lift vanished. Value that depends on contrast is not value — it is ranking, and ranking moves every time the room changes. https://bgodinspired.com/index.php/entertainment/human-made-vs-ai-why-the-maker-changes-the-value/
Questions People Are Asking
Do people really value human-made art more than AI art?
Yes, and the gap is large. In research from Columbia Business School using 28 images — half lesser-known works by artists including Paul Gauguin and Andy Warhol, half generated to resemble those styles — participants valued the art labelled AI-generated about 62 percent lower than the art labelled human-made. Notably, those same participants agreed the AI-generated pieces showed the same level of skill and nuance. They marked the AI work down specifically on creativity, on the amount of labour they assumed went into it, and on monetary value.
Does putting a “human-made” label on something increase its value?
Not reliably. A study of 618 participants published in Frontiers in Psychology in June 2026 by Han Sol Lim, Bong Gyou Lee, Yoon Hi Sung and Chang Won Jung found the effect was one-directional: the “AI-generated” label significantly reduced how much effort people believed went into a piece, while the “human-made” label produced no meaningful increase compared with content carrying no label at all. The researchers cautioned that this may reflect the current transitional stage of the technology and could change if machine-made content became dominant.
Why do people rate AI art lower even when the artwork is identical?
Research by Kobe Millet, Florian Buehler, Guanzhong Du and Michail Kokkoris, published in Computers in Human Behavior, ran four experiments with 1,708 participants and found the same artwork was preferred less when labelled AI-made. The mechanism was creativity: people judged the machine-labelled version less creative, which produced less awe, which lowered enjoyment. The bias was strongest among people who believe creativity is a uniquely human trait.
Is human-made work always seen as more valuable?
No. In the Columbia Business School research, human-made art was rated more valuable when it was displayed alongside AI-generated art, but that increase did not appear when the same human-made art was shown alongside other human-made art. The lift came from the contrast rather than from the work itself, which suggests the value being measured was comparative rather than intrinsic.
What is the human-made vs AI debate actually about?
At its core it is a question about where worth comes from. Across these studies the object does not change — the same image, the same craft, the same skill, granted by the participants themselves — while the value assigned to it moves sharply depending on who or what is believed to have made it. That means people are not really evaluating the work. They are evaluating whether there was an intention behind it.
