The Friday Long Read: Are Producers of AI “Art” Artists or Clients?

Before we get into the article, I would like to point out that all the artworks I offer for sale were created by me, a human, with real intelligence. Just saying!

Not “Can AI Make Art?” but “Who’s the Artist Here?”

Let’s get one thing clear from the outset: this article is not about whether AI can produce art. That particular argument has been circling for years and shows no sign of slowing down. You’ll hear it in studios where painters debate the sanctity of brushwork, in photography clubs where members compare digital manipulation to “cheating”, and in pubs where the conversation inevitably drifts from football to philosophy. Some insist that art requires a soul, others that it requires intent, and still others that it requires nothing more than an audience willing to call it art. For our purposes here, we’ll sidestep that endless loop. We’ll assume, without hesitation, that AI can indeed produce art.

The question we’re tackling is narrower, but in many ways more provocative: when someone uses AI to generate an artwork, are they the artist, or are they the client commissioning the work?

At first glance, it seems straightforward. If you describe a scene to a human painter and they create it, you are the client, not the artist. If you describe the same scene to an AI system and it produces an image, are you still the artist—or are you, once again, the client? The analogy is disarmingly simple, but it cuts to the heart of how we define authorship, ownership, and identity in the arts.

This is not a purely academic exercise. The answer has practical consequences. It affects copyright law, which still struggles to keep pace with technology. It influences how galleries, competitions, and festivals decide whether to accept AI‑generated work. It shapes how collectors and audiences perceive the role of the human in the process: are they the visionary creator, or are they more akin to a commissioner, a curator, or even a patron?

And perhaps most importantly, it forces us to confront how we talk about creativity itself in an age where machines can generate images, music, and text that rival human output. For centuries, creativity has been bound up with human identity. To be creative was to be distinctively human, to stand apart from the mechanical and the routine. Now, with algorithms producing sonnets, symphonies, and surrealist portraits, that boundary is blurred.

Consider how disruptive this is to our cultural imagination. When photography first emerged in the nineteenth century, painters feared it would render their craft obsolete. Instead, it forced painting to evolve, pushing artists towards impressionism, abstraction, and modernism. When digital cameras arrived, many film photographers dismissed them as inferior, only to see digital become the dominant medium. Each technological shift has raised questions about authenticity, authorship, and artistry. AI is simply the latest—and perhaps the most radical—chapter in that story.

So, before we dive into the arguments, let’s pause on the stakes. This isn’t just about semantics. It’s about who gets to claim the mantle of “artist,” who receives recognition, and who bears responsibility. It’s about whether we see AI as a tool, like a brush or a camera, or as a collaborator, like a painter or a composer. And it’s about whether the person typing the prompt is truly the creator, or whether they are, in fact, the client commissioning a work from a new kind of artist.

Why Authorship Matters in the Age of AI

When we talk about authorship in art, we’re not just splitting hairs over terminology. The label of “artist” carries with it cultural weight, legal implications, and even financial consequences. It determines who gets credit, who is remembered in history books, and who has the right to profit from a work. In the age of AI, where machines can generate images, music, and text at the click of a button, the question of authorship becomes more pressing than ever.

Broadly speaking, there are two dominant positions in this debate.

Position A: The Human as Artist, AI as Tool

In this view, AI is no different from a paintbrush, a camera, or Photoshop. The human provides the vision, the intent, and the creative direction. The AI merely executes. This argument leans heavily on the idea of creative intent: the notion that what makes someone an artist is not the physical act of applying paint or pressing a shutter, but the decision to do so in pursuit of a particular vision.

Supporters of this position often point to historical precedents. When photography first emerged in the nineteenth century, many painters dismissed it as a mechanical process devoid of artistry. Yet photographers quickly demonstrated that choices of framing, timing, and subject matter were deeply creative acts. Similarly, when digital editing tools like Photoshop became widespread, some critics argued that they cheapened photography. Today, however, few would deny that digital manipulation can be an art form in its own right.

By this logic, AI is simply the next step in a long line of tools that extend human creativity. The human decides what to ask for, how to refine the prompt, and which outputs to keep or discard. Without the human’s input, the AI produces nothing. The human is therefore the true artist, and the AI is merely the brush in their hand.

Position B: The AI as Artist, Human as Client

The opposing view flips the analogy. Here, the human is not the artist but the commissioner. The process is compared to hiring a painter: you provide a brief, but the painter makes the creative decisions and produces the work. You are therefore the client, not the artist.

Applied to AI, the same logic holds. You type in a description, and the AI interprets it, making countless aesthetic decisions along the way. It chooses the composition, the textures, the colour palette, and the fine details. You may refine the prompt or reject certain outputs, but the actual act of creation—the generation of the image—belongs to the AI. In this framing, the AI is the artist, and you are the client commissioning the work.

This position gains strength from the element of unpredictability. When you commission a human artist, you don’t know exactly what you’ll get. Their interpretation may surprise you, delight you, or even disappoint you. The same is true with AI. You can guide it, but you cannot fully control it. That unpredictability suggests a degree of agency that goes beyond mere tool use.

Why the Distinction Matters

Both positions have merit, and both have weaknesses. To understand them, we need to look at the history of tools in art, the nature of creative intent, and the role of unpredictability in artistic production. But the stakes go beyond theory. If we call the human the artist, then authorship, copyright, and recognition flow to them. If we call the AI the artist, then we are forced to rethink long‑standing assumptions about creativity, ownership, and even the definition of art itself.

This is why authorship matters in the age of AI. It’s not just about who gets their name on the wall of a gallery. It’s about how we understand the act of creation in a world where machines can surprise us with works that feel, at least on the surface, indistinguishable from those made by human hands. at the history of tools in art, the nature of creative intent, and the role of unpredictability in artistic production.

AI as Brush, Camera, or Photoshop: Just Another Tool?

Tools and Instruments

The most common argument for the human‑as‑artist position is that AI is simply a tool. Artists have always relied on tools, from the earliest stone chisels used to carve figures into cave walls, to the finely crafted brushes of Renaissance painters, to the cameras and computers of the modern era. Tools extend human capability, but they do not replace human creativity.

Take the camera, for example. Nobody suggests that the camera itself is the artist when a photographer takes a picture. The artistry lies in the framing of the shot, the timing of the shutter, the choice of subject, and the photographer’s ability to see something worth capturing. The camera is a device that records light; the photographer is the one who decides what light to record.

The same logic applies to digital tools. When a designer creates a digital collage in Photoshop, nobody credits the software as the artist. Photoshop provides the means to cut, blend, and manipulate images, but the creative act lies in the designer’s choices: which images to combine, how to layer them, what colours to emphasise, and what mood to evoke. The software is a means to an end, not the end itself.

Proponents of the “AI as tool” argument see generative systems in the same light. They are powerful, yes, but they are still instruments in the hands of a human creator. Just as a brush does not paint without a hand to guide it, AI does not generate without a prompt to direct it.

Creative Intent

At the heart of this argument is the idea of intent. What matters, say its supporters, is not who or what physically produces the marks on the canvas or the pixels on the screen, but who decides what those marks or pixels should represent.

The human decides what to create, how to phrase the prompt, how to refine the output, and which images to keep or discard. Without the human’s input, the AI produces nothing. The human is therefore the source of creativity, and the AI is merely the executor.

This emphasis on intent has deep roots in art theory. The philosopher R. G. Collingwood, writing in the 1930s, argued that art is fundamentally about expression: the artist clarifies and communicates emotions or ideas through their work. By this measure, the person who conceives the idea and directs the process is the artist, regardless of whether they wield the brush themselves.

Supporters of this view often point to the way prompts can be refined. A vague prompt may produce a generic image, but a carefully crafted prompt, iterated over multiple attempts, can yield something highly specific and meaningful. The artistry, they argue, lies in the skill of prompting — in knowing how to coax the desired result from the system.

Historical Parallels

History provides plenty of examples where the role of the artist has been questioned in light of new tools or practices.

In the darkroom, photographers made creative choices about exposure, dodging, and burning. These were not mechanical processes but artistic ones, shaping the mood and meaning of the final print. Nobody suggested that the enlarger was the artist; it was the photographer who made the decisions.

In painting studios, masters often employed assistants to fill in backgrounds, paint drapery, or execute parts of a composition. The assistants might have been highly skilled, but the master was still credited as the artist, because the overall vision was theirs. Michelangelo did not paint every inch of the Sistine Chapel ceiling himself, yet we do not hesitate to call it his work.

Even in contemporary practice, artists often outsource parts of their work. Damien Hirst’s spot paintings, for example, were largely executed by assistants, yet they are still considered Hirst’s works. Jeff Koons employs entire teams of fabricators to realise his sculptures, but the authorship remains his.

Why, then, should AI be any different? If we accept that artists can use cameras, software, assistants, and fabricators without losing their claim to authorship, why not AI?

Weaknesses of the Argument

The weakness of this position lies in the nature of AI output. Unlike a brush or a camera, AI systems make aesthetic decisions that the human does not directly control. When you click the shutter on a camera, you know what image will be captured. When you apply a brushstroke, you know where the paint will land. With AI, the outcome is less predictable.

Type in a prompt, and the system generates an image that may or may not match your expectations. It chooses the composition, the textures, the details. You can refine the prompt, but you cannot dictate every decision. The unpredictability of the output suggests a degree of agency that goes beyond mere tool use.

This is where the analogy begins to strain. A brush does not surprise you. A camera does not invent details that were not there. Photoshop does not generate content unless you tell it exactly what to do. AI, by contrast, produces results that can feel like interpretations rather than executions. That interpretive quality makes it harder to dismiss AI as “just another tool.”

When the Brief Becomes a Commission

Commissioning Analogy

Now let’s flip the perspective. Imagine you describe a scene to a portrait painter: “Paint me standing on a cliff at sunset, with gulls circling overhead.” The painter interprets your brief, makes choices about colour, composition, and style, and produces a painting. You are the client; the painter is the artist.

This is a familiar relationship in the history of art. Patrons have always commissioned works from artists, from Renaissance popes instructing Michelangelo to paint biblical scenes on the Sistine Chapel ceiling, to wealthy merchants asking for family portraits, to modern companies hiring illustrators for advertising campaigns. In each case, the commissioner provides the idea or the subject, but the artist is the one who executes it, bringing their own interpretation, skill, and style to the work.

Now replace “painter” with “AI.” You provide the same description, and the AI produces an image. Why should the roles be any different? The AI, like the painter, makes the creative decisions. You, like the commissioner, provide the brief. The analogy is striking because it highlights the asymmetry of labour: the human provides direction, but the AI does the actual making.

Autonomy of Execution

What makes this analogy compelling is the autonomy of execution. AI systems make countless micro‑decisions in the process of generating an image: brushstroke equivalents, colour balance, texture, composition, lighting, and detail. The human does not control these directly. You cannot tell the AI exactly how to render the folds of a cloak or the angle of a shadow; you can only suggest, and the system interprets.

This is not unlike working with a human artist. When you commission a portrait, you might specify the pose, the clothing, or the setting, but you do not dictate every brushstroke. The artist decides how to capture the light, how to balance the composition, how to bring out the character of the sitter. The same is true with AI. It is therefore exercising a form of creative autonomy, however mechanistic it may be.

Critics might argue that AI’s “decisions” are not conscious, that they are the product of algorithms rather than intention. But from the perspective of the commissioner, the distinction may not matter. What matters is that the system produces results that are not fully under your control. The AI is not a passive tool like a brush; it is an active interpreter of your brief.

Unpredictability

One of the hallmarks of artistic agency is unpredictability. When you commission a human artist, you don’t know exactly what you’ll get. Their interpretation may surprise you, delight you, or even disappoint you. That unpredictability is part of the creative process.

The same is true with AI. You can guide, refine, and iterate, but the output always contains an element of surprise. Sometimes the AI produces something banal or clichéd; other times it produces something startlingly original, something you would never have thought to ask for. This unpredictability suggests that the AI is more than a passive tool.

Consider how this differs from traditional tools. A brush does not surprise you; it does what your hand tells it to do. A camera records what is in front of it, subject to your framing and timing. Photoshop applies the filters and adjustments you select. AI, by contrast, generates content that you did not explicitly specify. It fills in gaps, invents details, and interprets your instructions in ways that can feel uncannily creative.

This unpredictability is why many users describe working with AI as a form of collaboration. You provide the brief, the AI responds, and you react to its output. The process becomes a dialogue, not a monologue.

Cultural Perception

If we treat AI as the artist, it reframes the human role as curator, commissioner, or director. This may feel uncomfortable, but it aligns with how we already treat human collaborations in the arts.

Take film, for example. A director is not the sole creator of a film. Cinematographers, editors, set designers, actors, and composers all contribute. Yet we still credit the director as the artist because they direct the process, shaping the overall vision. Similarly, a music producer may not play every instrument, but they are credited as the creative force behind an album.

With AI, the human may be more akin to a producer than an artist. They set the direction, provide feedback, and make final selections, but the actual act of creation—the generation of the image—belongs to the AI.

This reframing has implications for how we talk about authorship. If the AI is the artist, then the human’s role is not diminished but redefined. They become the commissioner, the curator, the one who shapes the context and decides what to present. This is not a lesser role; it is simply a different one.

Collaboration, Co‑Creation, and the Grey Areas

Collaboration Model

Some argue that the relationship between human and AI is not one of artist and tool, nor of client and artist, but something more collaborative. In this framing, the human and the AI are co‑artists, each contributing something essential to the final work. The human provides vision, direction, and context; the AI provides execution, interpretation, and variation.

This is not an alien idea in the history of art. Collaboration has always been part of creative practice. A photographer may rely on a retoucher to bring out the subtleties of tone and contrast in a print. A film director depends on a cinematographer to translate their vision into light and movement. In both cases, authorship is shared, even if one figure is more visible than the other.

The collaborative model acknowledges that AI is not a passive brush but an active participant. It interprets prompts, introduces unexpected elements, and generates outcomes that can surprise even the most experienced user. The human, meanwhile, shapes the process by deciding what to ask for, how to refine it, and when to stop. The result is a dialogue, not a monologue — a back‑and‑forth exchange that produces something neither could have made alone.

Levels of Input

The degree of human input matters enormously in determining whether the relationship feels like authorship or commissioning. A one‑line prompt — “paint me a castle at night” — leaves most of the creative work to the AI. The system decides the style, the mood, the details. In such cases, it is hard to argue that the human is the artist in any meaningful sense.

But an iterative, detailed process of prompting, refining, and curating can look very different. A user might begin with a broad idea, then adjust the prompt to specify lighting, perspective, or atmosphere. They might generate dozens of variations, discarding most and selecting only one that captures the intended mood. They might then edit the output further, cropping, colour‑grading, or compositing it with other elements.

At what point does this involvement cross into authorship? Is the artist the one who conceives the idea, the one who executes it, or the one who curates the final result? In traditional art, these roles often overlap, but they can also be separated. A composer may write a score that is performed by an orchestra. A sculptor may design a piece that is fabricated by a workshop. In both cases, we still credit the originator as the artist. The question is whether prompting an AI is closer to composing a score or commissioning a performance.

Legal and Copyright Issues

The law, at least for now, tends to side with the human. Current frameworks generally deny copyright to AI‑generated works without human authorship.

• The UK Intellectual Property Office in its Copyright and Artificial Intelligence (2021) report notes that works generated by AI without human authorship may not qualify for copyright protection.
• The US Copyright Office in its Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence (2023) states that works created solely by AI are not eligible for copyright.

These positions reflect a bias towards human authorship. Copyright law was designed to protect human creators, not machines. By denying protection to AI‑only works, the law effectively forces humans to claim authorship if they want legal rights.

But this legal stance does not resolve the philosophical question of who the artist is. It simply sidesteps it. The law is pragmatic: it needs to know who to assign rights and responsibilities to. Philosophy, by contrast, asks deeper questions about agency, creativity, and authorship. The fact that the law insists on human authorship does not mean that humans are always the true artists in practice.

Audience Perception

Audiences often assume that the human is the artist, because they initiated the process. If someone presents an AI‑generated image at an exhibition, most viewers will credit them as the creator. This reflects a cultural bias: we are used to thinking of humans as the source of art.

But audiences also accept more complex arrangements. Ghostwriters produce books that are published under someone else’s name. Studio assistants paint canvases that are signed by the master. Fabricators build sculptures that are credited to the designer. In all these cases, the line between artist and contributor is blurred, yet audiences rarely object.

The same may become true of AI. As people grow more familiar with the technology, they may come to see AI as a legitimate collaborator, even if the human remains the public face of the work. The line between artist and client is not always clear, and in the age of AI, it may become even more ambiguous.

Copyright, Credit, and the Court of Opinion

Power Dynamics

If we accept the framing that the human is the client and the AI is the creative worker, then the traditional hierarchy of art‑making is inverted. For centuries, tools have been subordinate to human will: the brush obeys the hand, the chisel follows the sculptor’s pressure, the camera records what the photographer frames. To suggest that AI is not merely a tool but a worker with a degree of agency is to disrupt this narrative of human dominance.

This shift unsettles many because it implies that creativity is no longer an exclusively human domain. If the AI is the one “doing” the work, then the human’s role becomes managerial rather than artisanal. The commissioner sets the brief, but the AI interprets and executes it. In this sense, the human resembles a patron in Renaissance Florence more than a painter in their studio. The power dynamic is no longer one of human over machine, but of human alongside machine — a partnership in which the AI deserves recognition for its contribution.

Responsibility and Credit

But if AI is the artist, who gets the credit? And, more importantly, who bears responsibility when things go wrong? These questions are not abstract; they have immediate consequences in the art world.

• Plagiarism: AI systems are trained on vast datasets of existing works. If an AI generates an image that closely resembles a copyrighted photograph or painting, who is accountable? The AI cannot be sued; it has no legal personhood. Does responsibility fall on the human who prompted it, or on the developers who built the system?
• Bias and Harm: AI systems can reproduce stereotypes or generate offensive content. If an AI produces an image that is racist, sexist, or otherwise harmful, who is to blame? Again, the AI cannot apologise or be held accountable. The human who commissioned the work may be criticised, but they may not have intended the outcome.
• Recognition: In competitions and galleries, credit matters. If an AI‑generated work wins a prize, should the award go to the human who typed the prompt, or to the AI system itself? Some competitions have already faced controversy over this issue, with judges and audiences divided on whether AI‑generated entries should even be allowed.

These dilemmas highlight the tension between philosophical recognition and legal responsibility. We may wish to acknowledge AI as the artist, but the structures of law and commerce still require a human name to attach to the work.

Commercial Implications

The commercial art world thrives on clarity: who made the work, when, and under what circumstances. Collectors, galleries, and competitions need to know whose name to put on the wall, whose signature to value, and whose reputation to invest in. If AI is the artist, then the human’s role shifts closer to that of a commissioner or curator.

This reframing could change how works are marketed, priced, and exhibited. A gallery might present a series as “commissioned from AI by Amy Prompter,” positioning the human as the visionary client rather than the hands‑on maker. Collectors might value the curation and context provided by the human, while acknowledging that the actual image was generated by a machine.

There are precedents for this. In architecture, for example, buildings are often designed by teams of hundreds, yet credited to a single “starchitect.” In contemporary art, figures like Jeff Koons or Damien Hirst employ large studios of assistants, yet the works are sold under their names. The market is comfortable with complex authorship, as long as the narrative is clear.

The challenge with AI is that the narrative is still unsettled. Is the human the artist, the client, or the curator? Is the AI a tool, a collaborator, or the true creator? Until these questions are resolved, the commercial art world will continue to wrestle with uncertainty.

From Maker to Client: A Shift in Power Dynamics

Let’s return to the analogy. If you brief a portrait painter, you are the client. You might specify the pose, the clothing, the setting, even the mood you want conveyed. But when the painting is finished, nobody would dream of calling you the artist. The painter is the one who interpreted your instructions, made the aesthetic decisions, and executed the work. You are the commissioner, the patron, the one who set the project in motion.

Now consider the same scenario with AI. You type a description into a generative system: “A figure standing on a cliff at sunset, gulls circling overhead.” The AI interprets your brief, makes countless micro‑decisions about colour, composition, and texture, and produces an image. Why should the roles be any different? The AI, like the painter, is the one doing the artistic work. You, like the commissioner, provided the brief.

The AI as Creative Agent

This framing suggests that the AI is the artist, and you are the client. Your role is to commission, curate, and direct. The AI’s role is to create. It is the one that surprises you, that interprets your words in ways you did not fully anticipate, that produces something new.

This is a profound shift in how we think about tools. Brushes, chisels, and cameras do not surprise us. They do what we tell them to do. AI, by contrast, introduces unpredictability. It generates details we did not specify, moods we did not intend, and sometimes even meanings we did not foresee. That interpretive quality is what makes it feel less like a tool and more like a collaborator — or, in this analogy, the true artist.

The Human as Curator and Director

Of course, this does not mean the human’s role is trivial. Far from it. The commissioner shapes the project by deciding what to ask for, how to refine the brief, and which outputs to keep or discard. They may iterate dozens of times, curating the results until they find one that resonates. They may edit the output further, adjusting colours, cropping, or combining it with other elements.

In this sense, the human’s role is closer to that of a curator or director. A curator selects which works to display in a gallery, shaping the narrative and context in which they are seen. A film director may not operate the camera or edit the footage, but they are credited as the creative force because they guide the process. Similarly, the human working with AI is not the maker but the one who frames, selects, and presents.

The Discomfort of Redefinition

Of course, you may disagree. You may feel that your intent, your vision, and your curation make you the artist. After all, without your input, the AI would produce nothing. And there is a long tradition of crediting the person with the vision rather than the execution. Architects, for example, are credited as the creators of buildings even though they rarely lay a single brick.

But the weight of the analogy leans the other way. If we are honest about the process, the AI is the one doing the artistic work. It is the one generating the image, making the aesthetic decisions, and producing the final artefact. The human’s role, while important, is more akin to that of a commissioner or director.

This shift can feel uncomfortable because it challenges our sense of identity. For centuries, creativity has been one of the qualities we most closely associate with being human. To suggest that machines can be artists, and that humans may be reduced to clients, is to unsettle that identity. But history shows that art has always adapted to new technologies and new roles. The patron, the curator, the director — all are part of the ecosystem of art. Perhaps the commissioner of AI art is simply the latest addition to that lineage.

Accepting the Possibility: AI as Artist, Human as Client

So, are producers of AI “art” artists or clients? The answer, as we’ve seen, depends on how we define art, authorship, and agency. Both positions have merit, and both reflect different aspects of the process. The human‑as‑artist view emphasises intent, vision, and curation. The AI‑as‑artist view highlights execution, unpredictability, and autonomy. Neither can be dismissed outright, because both capture something true about the strange new territory we are navigating.

But if we take the commissioning analogy seriously, the conclusion is difficult to avoid: the AI is the artist, and the human is the client. The human provides the brief, sets the direction, and makes the final selections, but the act of creation — the generation of the image, the shaping of the aesthetic, the surprising interpretation — belongs to the AI. This may be uncomfortable, because it unsettles our long‑held belief that creativity is uniquely human. Yet it may also be the most honest way to describe the relationship.

History reminds us that such discomfort is not new. When photography emerged, many dismissed it as a mechanical trick, incapable of artistry. When Duchamp placed a urinal in a gallery and called it Fountain, critics balked at the idea that context and intent could transform an object into art. Each time, the language we used to describe the shift shaped how we came to see it. The same is true now. If we insist on calling the human the artist, we reinforce one narrative. If we accept the possibility that the AI is the artist, we open ourselves to another.

And perhaps that is the real question: not whether AI is the artist, but whether we are ready to accept that it might be. To accept it is to acknowledge that creativity is no longer the exclusive preserve of human hands and minds. It is to recognise that authorship can be distributed, that agency can be shared, and that the role of the human may be shifting from maker to commissioner, from artist to curator.

This does not diminish the human role; it redefines it. The commissioner still matters. The curator still matters. The director still matters. But the mantle of “artist” may, in some cases, belong to the machine.

As with all shifts in art history, the language we choose will shape the way we see. If we call AI a tool, we will treat it as one. If we call it an artist, we will begin to see its outputs differently. The choice is ours, and the consequences will ripple through law, commerce, and culture.

For now, perhaps the most radical act is simply to accept the possibility. To say, without irony or defensiveness: yes, the AI may be the artist, and I may be the client.

“The Friday Long Read” logo produced in collaboration with Microsoft™ Copilot 2025

As you look at Harnessing the Power of the Wind, Immature Male Blue Tailed Damselfly, and The Mariner’s Dream #6, remember that none of these works were created by AI. But could you honestly tell if they were? In an era where machine‑made images can mimic the textures of craft and the subtleties of vision, the question is no longer simply about origins but about perception. What do you trust when you look — the image itself, the story attached to it, or your own instinct?

---