AI Art | Aesthetic Judgement

Like any new technology, AI art generators will affect different types of artists much differently. To speak about it as a monolith is misleading and unhelpful. To ban them outright is missing the point.

For someone who really wants an image and would normally hire an artist, they now have the opportunity to create versions of this image through an AI art generator. By inputting a string of written prompts the tool will produce several outputs which can then be iterated on through subsequent changes to the written prompts. This does not guarantee that the end result will adhere to the appropriate finished qualities which the user had in mind, but the quality of these tools in reproduction of various art styles and “realistic” representations is stunning.

The issues which are currently being raised around whether or not AI art generators are unethical and dangerous apply blame to the tool while failing to properly account for the intentions, skills and relationships brought by those who use them. My intention here is not to specifically fight for one side or another. Rather I would like to point out aspects of how AI art generators might disrupt the art world in ways which I have not yet seen discussed. I would also like to use this development in technology to highlight the relationship each of us has to art and media. There are important discussions to be regarding aesthetic judgements and the development of taste. Perhaps a different perspective on our relationship to tools such as this will be helpful in framing different conversations.

On Impacts to the Art World:

To speak about the world of Art as a unified field is highly dubious. The variety of ways in which visual media integrates into our lives is too varied to account for. From online ads, to branding to conceptual work, to video games and immersive experiences, to icons and navigation of our digital and embodied spaces, to visualization of information, and yes, to works of “fine art” intended for the gallery or museum, artists operate on a broad spectrum. Each with their own intention and relationship to both their tools and their final products. The following categories are certainly not exhaustive, but representative in the ways which AI art generators might have different sorts of impacts on the art world.

For many artist whose main income and focus is to generate images based on design briefs, this may well cause significant disruption to their industry. Concept artists, for example, are primarily impacted in many ways as this is specifically what they have been hired to do: turn concepts into images which do not yet exist. For these artists, crafting the images themselves may no longer be the primary focus of their job. Instead, one might imaging the concept artist needing to become an excellent interpreter of how the AI model receives textual prompts. Rather than skills in kit bashing and painterly rendering, the concept artist might need to become an expert at synonyms, historic artistic styles and vernacular relating to setting and mood. Given the extreme amount of examples which may be generated in a short period of time and the ease with which they can adjust and readjust the prompts, the concept artist will need to demonstrate excellent taste and refinement. No longer will it be a matter of generating “something” which captures the spirit, but rather the *exact* right look for the application. This exercising of aesthetic judgment will never be replaced by anything other than the creative mind which holds the unrealized vision up for comparison.

For illustrators working in commercial industries the challenge may be quite different. It is possible to imagine that some designers will be able to generate their own work without hiring out to an illustrator. Many others, however, will still need someone to provide them with collections to sell on fabric, home goods and other accessories. The market currently doesn’t have any particular favor for artists who create works digitally, through photography, or using paint and canvas. The value of the artist is that they have created a range of images which can be licensed regardless of initial medium. While many traditional artists working in the industry might bemoan the challenge of creating individual works themselves at a moderate pace, when another artist might leverage the power of AI generation to develop a broad swatch of designs in a sort amount of time. What of the artists working in watercolor who can only create so many images in a month as compared to their counterparts working digitally, who can create and edit multiple colorways simultaneously at much greater speed? Describing the unfairness of the tool in such terms would not be an addition to the conversation as it already exists. It has never just been about the efficiency of the tools. It is the ability to visualize the designs in the first place which sets these artists apart. Just because someone is holding the most expensive paintbrush, poised over a primed canvas and prepared with the finest oil pigments does not mean that they have the ability to generate an image others will appreciate, much less license. Even typing in “floral patterns” to a generator isn’t worth much. Choosing the best and most interesting, and being able to describe exactly what one is looking for requires artistic skills which are valuable. Whether an artist is able to submit one collection, or thirty, doesn’t ensure that their designs will be picked up.

Artists whose work is individualized as a personal expression may not see any threat. Indeed many artists have begun to explore the possibilities with which they can create new sorts of art with the aid, or even collaboration, of neural networks or deep learning systems. In more mundane cases there are also artists using image generators as a starting point from which to further develop their own works. Perhaps a bit like paint pouring or working from digital photographs, they don’t see the generator as a final product, merely a tool used to create the foundations for their own inspiration. For some artists who are posting works created through a generator, even if they do no other editing work afterwards to alter the returned image, this is still the creation of art. Knowing how to get a desired response out of an image-making tool through careful trial and error is still a craft of its own. Like paint pouring or other sorts of process art, there is a fine balance between controlling the inputs and being surprised by the outputs. The same could be argued about photographers. Just because they were in the right place at the right time to capture an amazing shot, does that mean that they are artists, or just lucky? Is it really different to talk about the understanding of a camera and the patience to wait for a subject in comparison to talking about how one learns to manipulate a generator? Certainly these conversations were had at the beginning of the twentieth century about those who “just snapped pictures” as opposed to being “real” artists. As in the examples above, the shift turns again on the importance of taste. Knowing what results are good and discerning where the real gems are amidst the hundreds or thousands of lackluster images.

On Intention:

The primary point I would like to focus on in this discussion revolves around intention of the artist. Is an artist someone who happened to create something which other people find value in as art, or is the artist necessarily someone who set about trying to create a work of art?

Is an artist hired based solely on their technical skills, or on their vision, or their ability to translate the abstract into a tangible end result?

Is the architectural draftsperson as much of an artist as the graphic designer who can adhere creatively to rigid specifications? Is the illustrator who can ink or color the pages of a graphic novel with consistency and style in the same category as the HVAC engineer who can adapt functional and elegant solutions to any convoluted home layout?

Are artists who bring their own inner worlds to life, advise other artists or outline future potentialities more artistic? How do we place style over content? Is it more artistic to represent something which has never before been grasped, or to present the familiar in a way which it has never before been experienced?

Do not misunderstand me. This is a rhetorical exercise and not a ranking of artistry. Depending on the final execution I am very happy to see the work of an HVAC technician as more artistic than that of an illustrator who is phoning it in. The sliding scale on which each artist falls does, however, impact their relationship to technology such as AI art generators. Those artists whose time, energy and employment rely upon technical execution, conceptual generation, layout or other technical aspects will be significantly impacted by this technology since it overlaps so wholly on top of these skills. Artists whose role relies upon their judgement, vision and narrative ability may see little impact, or even more opportunities.

As AI generators become ever better at filling roles which were traditionally focused more on technical skills and rendering, it will become less of a marketable skill that one can hand-render in perfect perspective.

On the other side, lowering the bar on ease of image creating to such a low level also creates the issue of determining what to create, in which way and for what purpose. Perhaps the creative teams will pivot to be more focused on the vision rather than the execution. Perhaps the intangible artistic skills will become more valued. Anyone can type prompts into a generator, but only someone with the right creative vision will be able to put together the concept with an appropriate manifestation.

On Theft, Intent, and Style:

The issue of artists’ works being fed into the data pools which help these generators learn new skills is certainly worth keeping at the forefront of discussion. Serious issues related to the intellectual property rights need to be addressed as we move quickly into a world where many such collisions will continue to challenge our current ways of thinking. It isn’t new that artist styles have been co-opted by others and essentially stolen. The pace, and ease, with which others can use AI to co-opt signature styles does create a uniquely challenging situation for independent artists. Theft is theft, whether the person using the AI generator understands that or not. What is unclear, however, is whether an artist should be able to stake any claim on their “style”. Where does one draw the line to determine whether another artist, whether painted by hand or through a digital tool, has copied someone else’s style as opposed to adapted it? If the crux of the argument comes down to intent, then perhaps it would be possible to enforce such a move as logging textual entries as they are input. If someone is including the names of specific artists when crafting their prompts, that could be a clear line demonstrating intent which could aid in enforcement. On the other hand, using the phrase “in the style of” shouldn’t always be seen as theft. How many musical artists have written their own works in the style of Bach, or Mozart or the Beatles? This sort of blending of inspirations is a cornerstone for how art develops and evolves. Artists, as humans, can only work with inputs they have already absorbed. Incremental development relies upon styles, techniques and content which has been previously integrated into the psyche of the artist. In this way the AI generators are not unlike the working mind of many artists.

To be caught up in a battle about what sort of art is being generated is to substitute any final representation of a style as if it were on equal footing.

Many have asserted that AI art has no “soul” as compared to a human artist. If that is the case, then where is the worry? If it is self-evident that this form of art is simply not as good, then we will automatically dismiss it when we see it. That is, of course, to confuse form as content, amidst a variety of other factors. Just because a work is created by a human doesn’t make it soulful, nor does every work created by a master of their media automatically a masterpiece. If I were to encounter an image which moved me, opened me up to new ways of looking at the world, and continued to live in my subconscious even after the encounter, I would be happy to think of it as art, no matter how it had come into being.

Artistic Judgement and Individual Taste:

Even writing the words “the cat is already out of the bag” is something which I don’t find particularly helpful, but I also believe that it is an important truth. In the long history of technology there have been vanishingly few times in which it was possible to close pandora’s box. Once something has been created and let out of the lab, there is simply no getting it to go away. Like social media, like cameras on our phones and the internet in our pockets, there is zero chance that AI created art is going to reverse course and diminish from culture.

The current pain points revolve around infringement of rights for independent artists due to the work of several bad actors. The technology certainly makes this easier, but hasn’t created the problem. Future concerns, I believe, are going to come as the technology begins to disrupt the many industries which make use of visual artists in a variety of capacities. In some cases this will allow for artists to relieve themselves of menial and unrewarding “artistic” tasks, while for others it will mean the end of a job, and for others still the opening of new frontiers not yet imagined.

Beyond all of these issues I see a larger disruption. Even in our world as it existed before this technology became widely available we have been drowning in visual content. Our ability to reproduce and display images has been increasing, and this technology removes certain roadblocks at the level of content creation. Any bottleneck for image creation that currently exists due to lack of available artists or simply production time is about to be dramatically reduced. Not only will this make content creation faster, but it will also allow content creation by those who have not been trained, or indeed, might not even be particularly interested in crafting a piece of quality. A similar analogy might be drawn to the recent rise of applications such as Canva, which tout their ability to make anyone a graphic designer.

This increase of the total volume of visual information and the increased frequency with which it scrolls through our feeds each day is a much more potent threat to our collective sense of aesthetic judgment. What counts as good, as passable, as eye-catching, as shocking, have already become diluted into a slurry of questionable value.

Developing a strong aesthetic judgment will be a key skill in this landscape. Not only for the artist but for anyone interested in consuming artistic media of quality. While there is much discussion about AI generated art has no “soul”, I think that we need to reconsider writing off an entire segment of visual culture based on how it was produced. The value of a work of art, or even of a working image, does not derive from any one area. Indeed it can be different for each viewer, and different within a specific context. This has always been the case, even if we tend to think of “good” taste in terms of what is shown in museums.

It is not interesting to look at the taste of a culture, especially since it seems very little value has been placed on this aspect of western culture within the last several decades. That is also why it is important to clarify what I mean by taste. Not that generalized ability to know what a society thinks is good, what the experts have deemed worth looking at, or what the critics understand to be important, but a personal sense of taste for each individual. Taste is the ability to discern for oneself what creates, sustains and nurtures value within an individual life. It is a continuously developing skill which grows through exposure to art and self-reflection. Unlike academic knowledge it cannot be picked up from a book or handed down from a teacher. It must be cultivated consciously through practice.

Simple questions may help the individual to think about what visual information they are consuming. Some questions like: what was my initial reaction to the work? Why did I choose to spend time looking at this particular work? Did it stop me in my tracks, or did I single it out for attention? Did it bring me a new perspective? Did it challenge me? How did it make me feel to experience it? Would I find it interesting to experience it again? What specific aspects of the work continue to stick in my memory, if any? Is it something I find myself wanting to share with others? Did it trigger any ideas or spark me to action in any way? Did it increase my connection to others, or make me feel more isolated? Did it make me want to learn more about the artist(s) who created it?

These sort of questions can help point out as much about us as about the works. Indeed, taste isn’t the development of an aesthetic framework which exists outisde of the viewer. Taste is developing a sense of our own values. It is a way of discovering more about ourselves, specifically what sort of art influences us and why we think that it is important.

Developing a strong sense of taste helps one to choose how to spend time in the world. Naturally, as taste develops, each person will spend more time around media, situations and people that foster their values, challenge them in healthy ways and add meaning to their lives. Conversely, aided with a strong sense of taste it is much easier to know what aspects of media or activities do not further, develop or challenge one. It becomes automatic to let those unhelpful and time-wasting aspects of life fall away.

The result of more people consciously developing their sense of taste will be compounded: First, the individuals themselves will be able to navigate a multifaceted and variegated world with confidence while maintaining connections to that which brings them meaning. The more individuals who develop and act on their sense of taste will ultimately shift the baseline for public media as a whole, for they will not wish to create media which does not meet their standards, and nor will they wish to consume it.

AI art generators are going to cause dramatic changes to the ways in which art is created and will have impacts for artists in a variety of ways. AI art, however, is not itself valueless or base or secondary just because of the tool that was used. It is simply more incumbent than ever that each of us spends time developing a sense of what our values are. Not just in art, but in all aspects of life. It is only the person who knows where and why to spend their time who will not be swept away by the torrents of cultural media which are already pouring down on top of us each day. Part of those values will naturally have to do with how we come to value artists themselves and what role we foresee them having in society. Where technology infringes on that role in illegal or discriminatory ways we must be ready to step in and create space. That likely doesn’t mean, however, the elimination or boycott of AI generated art. It requires a reorientation towards how we view, consume and value art and media of all kinds. The art that springs from a culture, the representation of that culture in artifacts, is not the “creation” of that culture, it is a reflection of it. We live in the world that reflects who we are. The only way to change that is to change ourselves.