What Human Creativity Means Now

The romantic conception of creativity — the artist as a unique individual vessel through which authentic feeling is transformed into form — was always partially mythological. Painters learned from other painters. Writers stole from other writers. Musicians played in traditions. The jazz improviser worked within a harmonic language developed by hundreds of predecessors. The romantic myth of original genius was useful for marketing albums and securing patronage, but it described very few actual creative acts.

What it did describe, accurately if imprecisely, was a real difference between a human creating something and a machine generating something. The difference was not one of pure originality — humans are not original in the way the myth suggests — but of something else: interiority, experience, the transformation of lived reality into form. A painting by Mark Rothko is valuable not because its color relationships are technically unique (they are not) but because those color relationships came from somewhere inside a specific person’s encounter with consciousness and are addressed to another specific person’s encounter with consciousness. The communication is between subjectivities.

AI has no subjectivity. This seems like the obvious place to anchor a definition of human creativity. It turns out to be less stable than it appears.

The Problem with Interiority as a Criterion

Interiority is not publicly verifiable. When you experience a piece of music as emotionally authentic, you are responding to the music, not to evidence of the composer’s inner life. The same is true of visual art, literature, and film. You infer interiority from the work; you do not directly observe it.

This was always true, but it was not a practical problem when the only things creating music were people. Now that AI can create music that produces the same emotional responses as human-created music — the same sense of being understood, moved, or challenged — the inference from response to interiority becomes unreliable. People are experiencing AI-generated music as emotionally resonant without knowing it was AI-generated, in exactly the same way they experienced human-created music. The music is doing the same work. The question of whether there was interiority behind it has become invisible.

There are two responses to this. One is to say: if the experience is identical, the interiority doesn’t matter. Value is in the experience, not in the biography of the creator. This leads to a kind of hedonic functionalism about art — if it works, it works — that has been explicitly or implicitly adopted by a large fraction of music consumers, who have happily listened to AI-generated music on streaming platforms without caring about its origin.

The other response is to say: the biography does matter, because what you are responding to in the art is your model of a human mind reaching toward you, and if there is no human mind, something essential about the transaction is fraudulent — even if you can’t detect it. This view leads to a strong position: knowing that a work was AI-generated should fundamentally change your relationship to it, regardless of whether you can hear or see the difference.

The Authenticity Shift

The market data from 2025-2027 suggests that a meaningful fraction of consumers have moved toward the second position — or rather, that “human-made” has become a quality signal that commands a premium in ways it previously did not need to, because it was previously the default.

The analog is organic food, which commands a premium not because most consumers can taste the difference in blind tests (the evidence is mixed) but because the provenance matters as an ethical and social statement. People pay more for organic food because of what it represents about production processes, not only because of what it contains. Similarly, people are beginning to pay for “human-made” creative content as a statement about creative processes, not only because the content is perceptibly different.

This is a real but limited mechanism. It works for premium markets. It does not stabilize the economics of the middle market, which cannot charge premium prices and has therefore lost market share to AI alternatives. And it requires legible authenticity — the ability to credibly certify that something was human-made — which turns out to be harder than it sounds.

What Certification Would Require

The most significant unsolved problem in the human creativity economy is the lack of a credible certification infrastructure.

An NFT-based provenance system that records the creation process of a work and attests to its human origin is technically feasible and has been implemented in several platforms. The problem is that recording the creation process doesn’t prevent someone from training an AI model to generate plausible creation-process recordings, which is a somewhat absurd regress but a real one.

Biometric approaches — recording the physical act of painting, the movements of a musician’s hands, the keystrokes of a writer — are more resistant to faking but raise obvious privacy concerns and do not cover the large fraction of creative work that is now conducted partly with AI tools as the human artist’s collaborators. A painting that involved AI-assisted color suggestions, AI-generated reference images, and a human who made all final decisions — is that human-created? The answer matters for certification purposes and there is no consensus.

Several collecting societies and guilds — the British Musicians’ Union, the Screen Actors Guild, the Association of American Illustrators — have proposed definitions that emphasize human creative control rather than pure human origin. The basic framework: if a human made the final creative decisions, exercised aesthetic judgment, and the work would not exist in its specific form without human choices, it qualifies as human-created regardless of what tools were used. This is analogous to how photography was eventually accepted as art — not as an imitation of painting but as its own form with its own standards of creative authorship.

The Survivor Bias in the Cultural Record

Here is something that gets insufficient attention: AI’s impact on creativity is primarily an impact on the cultural record, not on human creative experience.

People still paint, draw, write, compose, improvise, and dance. The fraction of humans who engage in creative activities for non-commercial reasons — because it is satisfying, because it is social, because it is how they make sense of experience — has not declined. If anything, the availability of AI tools has enabled more people to produce things that look polished enough to share, which has increased the volume of amateur creative production.

What AI has affected is what gets resourced, distributed, and paid for. The cultural record — the music that gets played at scale, the images that get seen at scale, the stories that get read at scale — is increasingly AI-influenced, while the majority of human creative activity happens outside the commercial distribution infrastructure, unseen and uncompensated as it always mostly was.

This is disorienting because the commercial record is the most visible part of the culture. The professional creator whose livelihood was derived from inserting human creative work into the commercial stream has been economically disrupted. The amateur creator whose satisfactions are intrinsic has not been disrupted at all — may in fact have been empowered.

The tragedy of the current moment is not that humans have stopped creating. It is that the economic infrastructure that allowed some humans to create full-time, professionally, is contracting, and the human creativity that remains is increasingly concentrated in the intrinsic-satisfaction sector where it does not produce income. That is a real loss for the culture, and it is not ameliorated by the observation that people still paint in their spare time.