AI and the Creative Act

Many people are worried that AI is so powerful and so skilled that it will replace human effort and labor. And it is true – it has reached capability levels where it has the skill and speed to outcompete humans at high-quality productivity, and it appears likely to only continue to get better. In a world where these tireless, high-skill capabilities threaten to radically displace human labor, it becomes important to look carefully for what cannot be replaced. There is a very specific lens that helps reveal three core capacities of human beings that AI is not able to replicate or perform in any way – and it is one that is easy to miss. We have to look at what AI is fundamentally unable to do in the creative act.

We can learn the most about this by looking at the creative process of an artist. It is noteworthy that across every genre, across many often controversial approaches to art – where an artist may issue instructions to a machine, or to a team of people, or may use software to produce their art, as well as very traditional artistic processes where the artist brings their own individual skill to every part of the production process – these insights about the creative process hold true.

Initiation

An artist has an individual biography. That biography informs a perspective on the world, and also frames certain capacities and limitations in their ability to express themselves – which often results in a style unique to them. It is within this backdrop that the first part of the creative process begins: the inspiration or impulse to make a piece. An author writes a new book. A painter begins a new painting. A musician authors a new song. This small act of starting something new is easy to overlook, but it is a profound creative act that no software – AI or otherwise – is able to do. Every astonishing creation that AI can produce required a human to issue a prompt describing the start of something. AI has near-infinite capability but zero initiative. It can build almost anything, but it can’t be moved to build something. It reacts to what a human proposes. It can’t originate from its own judgment about what would be worth making. We don’t know exactly why biography generates the felt necessity to originate – but it does. Something in experiencing life as an individuality, within a body, within a time and a place, makes a home for a creative spark. Where the impulse to make something new – rather than simply respond to requests – becomes possible. Where free will to create something inspired simply from within is possible. AI has none of this. It exists outside of time, outside of place, outside of any body that could be moved by the world. It has no interior from which something could emerge unbidden.

Sustaining

The second part of the creative process is the internal judgment the artist exercises as they work – about the quality and the success of their effort. This is a complex process, and one that requires long periods of producing work that is suboptimal. The artist may iterate on some part of a piece for a long time before arriving at an inner knowing that it is now right. They may put a piece aside and return to it over long periods, often producing other work in between. But something sticks – a sustained will to remain in an unproductive process, to keep attempting a piece that is showing no promise, to slog through the parts of the creative process that are clearly ineffectual. We don’t know exactly why genuine in-process judgment requires a situated self – but it seems a necessary condition. The will to continue in spite of failures and setbacks, or to pause and take up again a piece when inspiration returns, cannot be reduced to self-monitoring toward a goal. There is a deeper ember which burns, carrying the process of the creative act through the long journey of its making. Sometimes it is playful, sometimes determined, often frustrated. But there is something in the emotional life of the artist combined with their will that allows for the long birth of a creative piece – and this is quite different from a carefully calibrated feedback loop, from simulated annealing, from any process of self-tuning oriented toward an outcome. Those are optimization strategies. What the artist does is something else entirely: they live with the work, and the work lives in them, and neither is quite finished with the other until something inside says otherwise.

Completion

The final part of the creative process is knowing when a piece is done. Not when it has simply accomplished a task – AI is very good at that. It is more the inner knowing that arrives after a long process of working a piece, when the artist can step back and say: this is done. When the author stops revising and editing the book. When the painter steps away from the canvas satisfied. When the musician stops reworking the song. This is one of the great mysteries of being an artist – there is nothing self-evident in the process, nothing in the piece itself that announces its own completion. But it seems likely that the overall biographical arc of the artist strongly influences this sense of closure and judgment. A younger artist might consider something complete much sooner – having had less experience with a medium, less depth of self-expression, less tolerance for the long creative process. But with time the artist develops skill and capacity, often shifting the creative process itself. Fewer unproductive branches early on. A greater ability to go further with a piece before it arrives at a sense of closure. There is some interplay of skill, relationship to the creative process, and the initial inspiration for the piece – all of it situated within a biographical context that only deepens over a lifetime. AI can tell you when a task is complete. It can’t tell you when something is done. It has no biography against which completion becomes legible. No accumulated relationship to a medium, no evolving capacity for self-expression, no lifetime of prior pieces informing the felt sense of this one finally being whole.

The Life Behind It

What AI has absorbed is the will of effort – the skill, the execution, the sustained labor of craft. It can do this without fatigue and with a precision that exceeds most master craftsmen. That part of the creative process is no longer the bottleneck. But the parts that initiate, sustain, and close a work – that emerge from a life, a biography, a situatedness in time and place and social moment – remain entirely outside its reach. The work AI produces doesn’t belong to a life. And that absence isn’t a missing feature. It’s the whole thing. And so the answer to the anxiety at the beginning of this essay is not that humans will find ways to compete with AI at the level of skill and productivity – they won’t, and don’t need to. The human place in a world of powerful AI is the same place it has always been: at the origin of the creative act, inside the long struggle of making something, and present at the moment of completion. If anything, this moment calls for a revitalization of the arts – a renewed investment in the uniquely human capacity to start, sustain, and close a creative work. Because the labor was never the irreplaceable part. It was always the life behind it.