Joanna E. Kabala
Artistic Intent
This article presents reflections on human art and its relation to AI-generated art. It focuses on artistic intent and forms a critical response to generative AI art, highlighting the misleadingness of the anthropomorphic metaphor. Based on observations made while creating generative animations, which enabled an insight, to a certain extent, into how the creative process evolves, develops, and accelerates with the use of AI tools. The creation process began with an image, both hand-drawn and prompt-generated. It progressed through sequences of images, frame-by-frame animation, and ultimately to generative text-to-video rendering from a reference image.
While creating, it became evident that the starting point for any AI rendering originates—and likely always will—outside of computational AI systems, which need to be instructed to render anything. AI often excels at associating faster than humans can and occasionally produces entirely unexpected imagery. However, this does not prove that it possesses imagination or the ability to create meaningful images.
Artists can use it as far as their imagination allows; AI is a kind of intelligent paint. There is no doubt that technology will continue to advance, but will AI creations ever surpass human capabilities? It is doubtful because the barrier between human art and AI-generated art is not dictated by computational power or aesthetic principles of art, but by physical and mental properties that essentially differentiate humans from computational systems, even if some metaphorical convergences can be found. The starting point of the creation process emerges within a space largely unknown to us: our minds.
A few years ago, around 2022, the media and research worlds debated whether AI would transform our lives. At the beginning of 2025, this was no longer a question of if. But the question of how much, to what extent, it has already done it.
Art history teaches us that new technologies have always faced challenges in gaining acceptance, from oil paints in the Renaissance to photography and computer graphics. Nevertheless, AI art generators have seamlessly entered the scene. Creators have suddenly witnessed how a technology that used to require some time to be widely accepted in the creation process, so that the outcome of such a process could be called “art”, has transformed in a very short time into an “entity” perceived as a threat to artists. This shift is partly driven by the exponential growth of technology, which impacts all aspects of our lives, and partly by the overwhelming influence of media driven by corporate interests. This phenomenon poses a real social danger to artists, as it fosters the false belief that technology possesses a creative power surpassing human potential.
This belief is false because it stems from the anthropomorphic metaphor, a deeply ingrained concept in today’s world. While we must live with this metaphor, the assumption that machines resembling humans would make communication with technology easier remains a form of wishful thinking. Comparing machines to humans, or humans to machines, is a misconception. Our brain does not function like a computer. Even if it did, the process of thinking appears to take place in the mind, not the brain. And over the domain of mind, we can’t say that much at all.
While AI has profoundly changed our lives in all domains, including fine arts, there are no indications to suggest that AI creates in the same sense as humans do. What it does is high-quality visualizations, rendered based on the current status of its information space. However, AI does not create, so to speak, by itself because it possesses no self. There is no space in AI where an intent to create could emerge. Even if a program can be written that seemingly behaves like that, it will always remain just a simulation, even if close to a perfect one.
To explain the difference between human art and AI-generated art, it is valuable to distinguish between artistic intent and artistic inspiration.
Intent is the purpose behind the artwork. It holds the meaning of the artwork, it is a semantically grounded seed, a concept that, if developed, may take the form of an artwork.
Inspiration serves as the driving force behind the creative process, it can stem from anything, a mere coincidence (or Jungian synchronicity). Many concepts will remain undeveloped and will not take any shape if the inspiration to give a particular form is not found or stumbled upon.
Outside of art, “intent” is a term often used in psychology to describe the purpose, goal, and meaning that drive our decisions and actions. Intent is something that precedes action and influences behavior. It is a mental state that, in the context of art, serves as the motivation driving artists to convey messages through their works. These messages do not need to be clear statements. They may take the form of emotions, personal experiences, or memories. Works of art are expressions of these individual mental states.
The distinction between intent and inspiration, which helps to grasp the difference between art created by humans and AI-generated art, emerged from the assumption that art is an expression of human consciousness. Consciousness is perceived as a realm or force that is inaccessible to AI. Individual and collective consciousness are understood here with all prefixes – unconsciousness, subconsciousness, supra-consciousness – all these states of mind that make us feel human and differentiate us from the rest of the world, especially from the world of machines.
Creative people are aware that they create, even if ideas are popping out of their minds uncontrolled, out of nothing, and they don’t know why some ideas stay longer, until they are developed into artworks. However, because they are aware, they know that the power that pushes them to create cannot be easily described and transferred to another being, even if it is also a conscious being.
The artistic intent is goal oriented. The approach is often planned, so some careful thinking is performed before making a decision or taking action. Artistic intent is an analytical deliberation of various options, weighing pros and cons, and evaluating potential outcomes. From selecting the technique of initial sketch drawings to culture studies fueling the artworks’ semantics. Up to the selection/rejection process, when looking at the vast outcome of the AI art generators, as not every or any AI rendering will do when an artist intends to convey a message.
On the contrary, the inspiration is spontaneous. It is an impulse, a sudden and often unplanned urge to act. Actions driven by impulse may occur without an extensive prior thought or intention. The approach is unplanned and intuitive. It occurs in immediate reaction to the flow of ideas. Spontaneous actions often result in unexpected outcomes, as they are free from the constraints of planning. Then the result is frequently an untitled expression of a gesture, as if the hands were drawing while the mind is temporarily switched off. Such bursts of creativity essentially fuel the creative process, keeping it moving forward.
Intent and inspiration are both essential components of artistic creation. Artistic intent is about what, the content of deliberations that keep the artist’s mind busy, while inspiration is about how, about getting an impulse to create and give shape to ideas.
This is why artists who carry some artistic intent often seek inspiration to create a work of art, which may take different forms depending on the source of inspiration. Artistic intent arises in the mind, and it may fluctuate under the influence of a variety of inspirations.
At this point, the anthropomorphic metaphor could be applied. If anthropomorphism helps some people to make sense out of technology, there is no need to reject it completely and immediately. Although it is good to keep in mind that it is misleading, it makes people believe that AI is something, which it is not.
However, this metaphor may be used to explain that the prompt becomes a source of inspiration to an AI art generator. Although it is good to remember that this is always the human who had an intent, and AI only computed an answer to the prompt.
It did the computational work quickly, sometimes on topic, sometimes completely off. Certain divergences in AI renderings may be seen as useful, even when they omit parts of the prompt or unexpectedly introduce objects and figures not originally included in the prompt. Ultimately, it is the human designer who determines whether variations and alterations are useful at all.
The prompt initiates a loose stream of AI associations that exceed humans in the tempo of pattern recognition, but by no means is human-like intelligence, which constantly constructs ever-changing semantic structures. AI is blind to semantics. It does not see what it renders. It is not aware that the pictures it creates may sometimes make no sense. It even produces semantic coincidences. How otherwise name such renderings: (fragments of) chairs, when a table has been prompted, a tail apart from the animal, three fingers instead of two in the victory sign, or a telescope put against the mouth instead of the eye of an astronomer?
Skipping the anthropomorphic metaphor leaves us with what generative AI is in its essence. Speaking colloquially, it is not much more than LLMs, large language models, which are still better and better in retrieving and associating information based on prompts, far faster than humans can, making them essential computational aids in our rapidly accelerating reality.
System developers may argue that AI art generators are not technically the same as LLM. However, seeing it as a large model seems much more appropriate than seeing it as a sentient being.
Despite the actual technology behind them, there is a certain feature that all generative systems have in common. It puts additional constraints on AI creations (next to the lack of artistic intent resulting from the lack of a mind).
Generative systems produce chaos; this is their very nature. It is easy to notice during longer rendering sessions, and could be better explained through another creative field, namely music. Music, due to its mathematical foundation, seems even closer to being taken over by AI than visual art is. However, just like in generative AI art, in generative music, there must also be some human thought at the very beginning. And later in the process, to make sense out of a vast number of generative samples that are catchy and well-defined stylistically, it is still the human ear, combined with artistic intent, that is necessary to create a coherent whole. Otherwise, pure AI samples or riffs quickly become too similar to each other or, in other words, boring. This audible boredom, in visual terms, takes the shape of chaos, which could be described as drifting into unnecessary details.
The chaotic results of renderings in long sessions are becoming more alike rather than more precise. Unlike artists, who refine their precision as they develop their style, AI is becoming lost in details. While repetitions in human creations result in distinctive forms, the signature gesture that reflects artistic identity, the repetitions in AI creations lead to greater chaos.
To summarize, the concept that emerges in the human mind is the seed of artistic intent that may be developed into an artwork. Concepts cannot originate within AI, as it lacks a mind. Humans don’t know what the mind is, or where it is. We can only feel that it is something that allows for expressions of consciousness. The mind, as well as consciousness, may be external to the human body, which more and more contemporary scientists, philosophers, and other thinkers admit.
Artistic intent emerges from consciousness in real-time, shaped by individual experiences. It reflects the struggle of refining vague ideas into an artwork, a creation that acquires meaning and form, perceptible to others. It may or may not receive a title.
Programmable tools can be efficiently used to express artistic intent and may even serve as a source of inspiration during the creative process; however, they always remain just tools.
Without human consciousness, these systems are not capable of recognizing meaning. They render randomly, and only the direction from a human creator may steer the renderings towards symbolically meaningful pictures.
Also, AI does not intentionally seek inspiration. It has no apparatus for intent, so it merely awaits prompts, cleverly utilizing feeds and learning in the process. It can “talk”, reacts quickly, seems to “know” so much, but all its products are compilations, merely imitations, although sometimes very appealing, or even seeming perfect. AI continues to improve at performing ‘as if’—perfecting the art of imitation. It is good at pretending, if we stick to the anthropomorphic metaphor.
AI bears only a slight resemblance to humans, and these similarities do not benefit either it or us. It is human-like in how it sometimes doesn’t understand anything. Doesn’t matter the language used. Doesn’t matter if it was fed with the artist’s images. Sometimes, it doesn’t get the point out of the prompt.
Metaphors are essential when names have yet to emerge. They may stay forever, then words gain double meaning. As it happened with Engelbert’s metaphor of ‘mouse’. But this metaphor was at least abstract or even surreal, ensuring it was not confused with its real-world counterpart. Unfortunately, the anthropomorphic metaphor of a computer system is fundamentally misleading. It constructs a skewed perspective of reality that may become a trap for many. And any claim of sentience in computable matter only blurs the reality more, forming an essential drawback to the collective (un-)consciousness, which may ultimately be dangerous for humanity, indeed.
The final observation in this article is important for the arts, mainly the visual arts.
An attempt to communicate with an AI generator using only an image failed. It required at least four words to initiate rendering. This may support the conceptual art postulate that concepts hold supremacy over visualization and other forms of expression. At the very least, it is an appropriate example that provides a clear explanation for why generative art is classified within the post-conceptual stream in art history. On the other hand, this is another example that AI art generators are no more than just tools. Humans project concepts straight from their minds, and the form of an artwork will differ depending on the tools used. A computational system will not produce concepts, but it will rather await a written instruction to render an image or other art form to which it was specifically created.
Joanna E. Kabala, Eindhoven 31 May 2025
https://joannakabala.nl/
References
Federico Faggin. Irreducible. Essentia Books, an imprint of John Hunt Publishing. 2024
Mark Johnson. The Body in the Mind. The University of Chicago. 1987
Ray Kurzweil. The Age of Spiritual Machines. Viking. 1999
Sol LeWitt. Paragraphs on Conceptual Art. Art Forum. 1967
Tor Nørretranders. The User Illusion: Cutting Consciousness Down to Size. Viking. 1998
Penrose, Faggin & Kastrup. Quantum Consciousness Debate: Does the Wave Function Actually Exist? https://www.youtube.com/watch?v=0nOtLj8UYCw&t=992s. Essentia Foundation. 2024