Please view the PDF for images and bibliography.
The rise and use of Artificial Intelligence (AI) is a defining topic of the early 2020’s, with popular beliefs about its trajectory finding themselves polarised. As we grapple with the implications of AI across various industries, its inclusion in contemporary art practices has posed new questions about the intersection of AI and artists. This essay will explore the role of new media, particularly generative images, in contemporary art. It will determine whether AI-produced artwork can be considered part of an authentic practice, and analyse the significance of intent in determining its validity. There are two questions regarding authenticity in art. One being the work’s integrity as a piece of art, regarding the legitimacy of the creative process. The other concerns the authenticity of the artist’s expression itself. Through this analysis the essay will aim to provide insight into the evolving relationship between technology and artistic practices in the current era.
Intent and authenticity of the practice are key determining factors in evaluating contemporary art.
Authenticity is understood as a genuine approach to oneself and others, being credible and aligned with one’s own beliefs and values. Inauthenticity, on the other hand, socially functions as a moral transgression (Gino, Kouchaki and Galinsky, 2015), inciting feelings of disgust towards oneself or others. The lines between the two are blurred and open to subjectivity. In the context of art, authenticity could be understood as the alignment of the artist’s self-concept and beliefs with their work. Consequently, an inauthentic artist creating inauthentic work results in an authentic artistic practice. In those cases where an artist’s practice isn’t an authentic expression of themselves, but rather a wider commentary, the waters become muddied. Rousseau theorised that “authenticity is derived from the natural self, whereas inauthenticity is a result of external influences”(Yacobi, 2012). This means that an artist’s inherent drive to create is authentic, but their practice, due to the nature of making work in a contemporary environment, can never be. The authenticity of a work – not in the sense of its legitimacy as ‘the real deal’, but in its genuineness – is an integral aspect of critiquing and understanding artworks.
To determine why AI cannot express authenticity, we first have to understand how it works. AI image generators use machine learning algorithms to learn from large datasets of images, which can include anything from paintings and photographs to 3D models and game assets. These training images should be as diverse and representative as possible, so that the AI generator can learn from as many various patterns and features. During training, neural networks identify and extract specific features from images, such as shapes, textures, and colours.[1] The machine can then generate new images based on input parameters, or prompts. An AI image generator does not have intentions of its own or the capacity for morality, which is necessary to determine authenticity. Neither can it reflect on said intent to express itself, as it is based on pre-existing images and will inevitably create biases through the selection of source material.
Another reason for the skepticism towards AI as a medium is that it is unchartered territory. Almost all media with “the exception […] of certain performing arts” (Benthall, 1972) utilise technology in some form. Media such as painting are “technologies which have been absorbed into traditional art” and typically “allow the artist much greater control over their activity” (Benthall, 1972) than new media. However, generative image models are faced with the inverse problem of having an exceptionally low barrier for entry in regard to skill and equipment. There currently aren’t any “established styles and conventions […] [that] are recognised” (Benthall, 1972) for their use, which is why it is integral to acknowledge the changes in conventions of the recent epochs, which inform the current boundaries set in place.
[2]Late modernity introduced ready-mades with Duchamp’s installations, such as ‘Bicycle Wheel’ (1913) [IMG 1] and the famous ‘Fountain’ in 1917 [IMG 2]. It sparked a heavy debate around the essence of art. Duchamp’s intentions were to challenge the use of appropriated objects as art, and in his drive to deconstruct art, he took the artist’s authorship with it.[3] His works “combined chance and choice, the arbitrary and the given” (Foster et al., 2016), they ‘”put conventional notions of art and artists alike into radical doubt; they were ‘works of art without an artist to make them’”. Through his proxy Beatrice Wood, Duchamp remarked in ‘The Blind Man’, that whether it was made “with his own hands […] has no importance [because] he chose it”. Duchamp intentionally removed his person from his practice, channelling his frustrations with the contemporary art scene of the early 20th century into work that at first glance seems illegitimate, a crude joke at the expense of himself. Despite Duchamp’s insistence on aliases, removing himself as the author, there is no doubt in the current view of his practice, that his work is an authentic expression of himself and that he retains authorship of it. He was marking a pivotal moment in art history, where the artist and the art began to be deconstructed and questioned.
Postmodernist pioneers, like Jeff Koons and Damien Hirst, continued Duchamp’s deconstruction of the artist as the maker of the artwork by delegating the physical making of the work to fabricators.[4] Koons argues for the authorship of his work through “having the idea or vision for the work” (Gunderson, 2023). [5] Many net artists feel a strong “connection to the work of [..] Duchamp and to the participants in Dada [..], all of whom helped to shift art practices away from traditional forms of pictorial representation” (Greene, 2014) through late modernism’s concept of “art as a process of nomination by the authority of the artist” (Foster et al., 2016).
Ready-mades are comparable to AI generated images in that regard. Therefore, critiques of AI-generated art follow a similar argument, with the addition of general concerns about the implications of intelligent, autonomous computer programmes.
Since the modernist period, “art has made itself the subject of itself, questioning its own legitimacy and that of the people creating it” (Bahr, 2023). Petra Bahr, member of the German Ethics Committee, noted in an interview with Art Magazin that art “may be less tempted to fall into extremes: neither into transhumanist dreams of superior machines leading lacking humans into a brighter future, nor painting the nightmare of self-creating systems declaring war against their creators on the wall” (Bahr, 2023).
Yet, these are the main themes of discussions around artificial intelligence in contemporary art, as shown in ‘Sinofuturism’ (1839-2046 AD) by Lawrence Lek, and will remain so until one theory proves true.
’Sinofuturism’ [IMG 5] is a video essay that discusses a future which is “embedded into a trillion industrial products, a billion individuals, and a million veiled narratives”, culminating in “multiple overlapping flows” ( Teixeira Pinto, 2020).
The essay is a catalyst for possible new net art movements, which explore technology and mass production under capitalism. The “rapid urbanisation and modernisation” of East Asian countries with the increase of advanced “technology [can] be seen as the site of a double alienation” (Teixeira Pinto, 2020), and is a globally prevalent concern. Similar themes in art practices were identified by Ana Teixeira Pinto in her essay ‘Alien Nations’ (2019), where she lays out different culturally informed visions of our future such as ‘gulf-futurism’ and ‘afrofuturism’.
Technology is viewed and used as a metaphor for “the new and foreign” (Teixeira Pinto, 2020). In Jung Yoonsuk’s “Lash” (2017) video installation [IMG 6], mechanical and sexual differences are highlighted through impossibly perfect sex robots being produced in a run-down factory.
While not touching on the relation between human and artificial intelligence specifically, it shows contemporary artists’ view of machines. They are at the forefront of human lives, both hopelessly dependent on one another as demonstrated by Lek’s chosen medium of ‘new media’, “incorporating unexpected technologies, such as video-game software and CGI animation, to create digital landscapes” (Network, 2022).
An integral aspect of artificial images is the ‘uncanny valley’ effect, an emotional response of unease, disgust and even fear when confronted with an image that has a human-like appearance yet isn’t human. The ‘valley’ refers to people experiencing an empathetic response to human-like objects up until a point right before perfectly emulating a person, where the positive response suddenly drops to being a negative one [IMG 7]. [6]
The concept was first coined by Japanese roboticist Masahiro Mori, who described “the phenomenon as ‘bukimi no tani gensho’, meaning “valley of eeriness” (Cherry, 2020).[7] Originally intended to inform the design of prosthetics, it has gained traction with people experiencing the phenomenon with artificial intelligence robots. Hanson Robotics’ AI android ‘Sophia’ [IMG 8] and Meta’s chatbot ‘Billie’ [IMG 9], which uses the likeness of Kendall Jenner, are two notable examples.
The human reaction of discomfort reflects the inability of the “non-human entity […] to measure up to the standards of an actual human being”, in other words to authentically copy a real human (Saygin, 2011; MacDorman & Ishiguro, 2006). In the same way that inauthenticity is perceived as a moral transgression in humans, the robot’’s inability to replicate a human evokes the same feelings of disgust towards it. With AI-generated art, the uncanny valley effect might be amplified by peoples’ fear of it becoming a threat to art making and creatives’ livelihoods.
The long nineteenth century brought up new perspectives on human identity, which AI confronts us with again by posing “the fourth great insult” (Bahr, 2023) to humanity, as the “threat to human uniqueness could be construed as a push to redefine humanness” (MacDorman and Entezari).[8]
This question of human identity and artificial intelligence has been prevalent in recent discourse, often in conjunction with commentary on the ‘uncanny valley’. One example is the exhibition ‘Uncanny Valley: Being Human in the Age of AI’ at the De Young Museum in San Francisco, which included artists such as Zach Blas [IMG 10] and Lynn Hershman Leeson [IMG 11].
The three art film series ‘Delphi Demons’ (2023) [IMG 12] by Kennedy+Swan, follows this question in a “humorous exploration of how AI creates images out of probability and variables, defamiliarizing the familiar”, by anthropomorphising AI.
In a scene in the video, a woman meets an oracle in the middle of a desert which answers in distorted images. This oracle, a re-imagination of Phythia the priestess of the temple of Apollo at Delphi, is named Wombo, after the AI program that created the images [IMG 13]. It “entices [viewers] to investigative views and lines of questioning in search of the construction mechanisms and even hidden messages” (Ulrich, 2023) only to reach the conclusion that the AI oracle is nonsensical.
Kennedy+Swan’s work often touches on the real vs fake and how to identify differences between humans and robots. It plays with the ‘uncanny valley’ effect to show us digital dystopias and utopias by merging “a variety of animation techniques” (www.newpractice.net, 2023) to make their films.
Kennedy+Swan keep control over their art by employing many mediums and using AI as a tool, when necessary, in order to highlight its characteristics within their imagined worlds. This also answers part of the question this essay poses: AI cannot make art on its own, it requires a human process of selection, interpretation and refinement. It is meaningless until we give it meaning.
Nils Pooker, a German painter, has gained traction for his use of new technologies in the process of making his work.
Another example of AI being anthropomorphized is in the painting “Art versus AI art” (2023) [IMG 14], in which Pooker explores the empathetic connection between humans and technology before the point of the ‘uncanny valley’. Pooker generated images of a personification of the AI model as the artist, with “the best result show[ing] an Android-like creature that […] holds a framed painting” (Nils Pooker, n.d.). A while later, he generated an image with the content “cute artificial intelligence” in order to address the ”anthropomorphisation of the AI tools themselves as supposed artists” (Nils Pooker, n.d.).
In his AI series, the artist generates images, pixelates them and paints the pixel image, which diffuses the image to make it unrecognisable. [9]
In ‘Landschaft’ [IMG 15] you can just about make out the landscape image, reminiscent of a painting in the computer game Minecraft, which were all “based on real-life paintings by Zetterstrand [IMG 16], who also did the work of turning them into pixel versions” (Geere, 2019). They share the use of a digital visual language in a traditional medium, e.g. oil painting. It creates a computer-generated look that appears more authentically fake than the generated ‘original’. The long, meticulous process of painting perfect squares negates all the advantages that AI provides for Pooker; neither the detailed generations nor the time benefits survive the process. His prompts include styles of different artists and themes from Greek mythology, knowing that the AI would not have any pre-existing images to copy. This, in combination with the further abstraction of the image and translation of the work demonstrates Pookers understanding of the tool and the intentionality behind his use of it. Through the process, he successfully claims it as a tool within his authentic artistic practice.[10]
AI-generated content has been infiltrating all parts of life, from celebrity clones on social media to generated images fooling unsuspecting facebook users.
One recent scandal of AI being used ‘inauthentically’ is the case of ‘Willy’s Chocolate Experience’ in Glasgow, Scotland, in February 2024. It was an “unlicensed event based on Charlie and the Chocolate Factory, which was promoted using AI-generated images and text” (Wikipedia Contributors, 2024) [IMG 17]. These were not representative of the actual event and caused a globally talked-about scandal, as parents showed up with their children and were greeted with a sad shadow of their expectations. The original website and marketing materials can still be found and they ” highlight [..] exactly what happens when one uses AI to generate pictures and text, without bothering to edit what it spits out” (Nolan, 2024).
With the accessibility and success of generative models, there have been cases of famous artists who used AI and passed it off as their own original creation. One such case is that of Emanuele Boffa, an Italian photographer whose work had previously been published in Vogue magazine. AI-generated images posted on his Instagram account were not marked as such, which caused backlash against Boffa, who describes himself as a “digital artist” (Chen, 2023) and photographer [IMG 18]. He explained: “I use all kinds of technology, […] but I don’t only use AI to generate my photos, I use Blender, Photoshop, and other systems” (Growcoot, 2023). Arguably, the public outrage over his artificial images was not about the use of modern technologies itself, but the lack of transparency. In response to an AI-generated image winning a prize at the Sony World Photography Awards, the author, Boris Eldagsen, rejected the prize, and questions of cross-media plagiarism were brought to the forefront [IMG 19]. Eldagsen stated: “AI images and photography should not compete in an award like this. They are different entities. AI is not photography” (Mouriquad, 2023).
Boffa’s generative images are nearly indistinguishable from real photos and his public image is first and foremost that of a photographer, the images “duped his followers into believing his portraiture was real when it was actually AI” (Growcoot, 2023). It is this avoidance, almost denial of the truth, that is an inauthentic representation of his practice as a digital artist and causes concern about the use of artificial intelligence.
As a reflection of our zeitgeist, contemporary fine art is faced with the impossible task of finding a legitimate space for AI to exist, one that doesn’t delegitimise the work it helped make.
For example, Charlie Engman, an American photographer, released a series of AI images which were just that – images that were generated using an AI system. They remained entirely unedited, a kitsch parody of the American Dream. These works might give insight into the biases AI inevitably perpetuates but display a lack of artistic intent to transcend their function as an exploration of the medium’s capabilities [IMG 20]. His use of AI is described as “[utilising] the technology’s limited reading of the inner states that connect to physical expression” (McClean, 2023), but fails to describe through which processes he pulls the work into authentically representing the human condition. Unless his intentions are to show a machine’s inability to be human, the lack of reclamation by the human artist means that his work doesn’t achieve what it sets out to do.
At first glance, the process American artist Jess MacCormack has adopted for her AI creations is exactly the same process used by Engman; a model is selected (in her case DALL-E 2), a prompt entered and the image information (e.g. size or detail) adjusted. This leaves the specifics of the image up to the programme, making room for the biases from the source materials to manifest.
In Jess MacCormack’s, ‘Your Trauma is Showing’ (2023) [IMG 21], “[…] [I]t isn’t clear anymore which parts are body and which are innards spilled inside-out.” The AI distorts bodies, opening the work up to interpretations about the effects of dissociation and trauma while attempting to free itself from traditional understandings of gender. “[T]he surreal is understood as a politically emancipatory concept in the use of AI. […] [A]rtists have been quick to use the new technological options to develop images of bodies which are removed from normative presentations of sex and identity” (Ulrich, 2023).
She intentionally relinquishes control over the work to the AI, utilising the “limitations to what this technology can produce” and likens them to “barrier mechanisms” that people develop through trauma (MacCormack, n.d.), demonstrating an informed reclamation of the ’AI’s nature.
Michael Borowski is an artist who uses AI to explore queer identities. He reviews the impact of discrimination against queer people in ‘the archive’ by creating scenes from a fictional bathhouse [IMG 22]. He creates salt prints with negatives derived from AI images in order to imagine scenes of places that “allow forbidden forms of sexuality to be secretly practised” (Ulrich, 2023). It reveals the AI’s potential to highlight “restrictions, which queer and homosexual life forms have long been under the mercy of“ (Ulrich, 2023), while further evolving artificial generations into faking missing evidence of the past.
MacCormack and Borowski aren’t simply technology users who input a prompt into an image generator and attempt to pass it off as an authentic piece of artistic work. What distinguishes them is the way they engage with the medium of AI. As described in the book ‘Science and Technology in Art ’Today’, “an artist, like other workers, has to organize his resources [in a way] that coordinat[es] technical resources with psychological or spiritual resources, so that there is a continuous process from idea to technical expression and back to idea” (Benthall, 1972). Through this process of evolving the work, both technically and academically, the artist is able to create something impactful.
Not just in art, but in all industries and areas of life, AI has demonstrated its potential to catapult the digital revolution into new dimensions. This poses ethical considerations and challenges in a fine art context. One of these is the question of ownership and authorship.
AI-generated images, as previously touched upon, are new images created from vast databases, meaning that they can be considered an amalgamation of the world’s entire image culture. The datasets often include the work of living contemporary artists and enable users to infringe on their copyright by generating work in their exact style. Famously, Greg Rutowski, a digital artist known for his fantasy landscapes and characters, reportedly had his work “copied more times than Picasso” (John and Hutchinson, 2023) [IMG 23].
Rutowski campaigned against the inclusion of living artists in AI databases, and in June 2023 won a court case against stable diffusion, permanently removing “the ability of emulating the style of specific aritsts” (Lanz, 2023) from Stability AI.. AI generators are easy and accessible, allowing any user to create copies of any included artist. Previously, plagiarising work was confined to a small group of people with the resources and/or skill to recreate it. This meant that identifying the violation of copyright and holding persons accountable was much simpler (e.g. Koons’ plagiarism of ‘Fait d’Hiver’ [IMG 3 + IMG 4]). Unfortunately, the responsibility of ethical use currently still lies with the users of the programmes, a fact that is demonstrated by the continued misuse of Rutowski’s work in a model “trained to recreate [his] style [which is] freely available on Civitai” (Lanz. 2023).
AI is often used as a tool to display transhumanist phantasies of a new age as well as to explore the intangible lived experience of the human condition. The two seem contradictory at first, but ultimately display the attempt of artists to place new technologies in context of themselves and speculate about their meaning, ranging from MacCormack’s exploration of what it feels like to live with dissociative identity disorder, to Kennedy+Swan’s videos in which humans give up control of their bodies to AI organisms.
On theme with the subject of this essay, I asked ChatGPT to collaborate by writing its own conclusion on this topic. Here is what it had to say:
“In conclusion, the appropriate use of AI as an art medium necessitates a nuanced understanding of its capabilities, limitations, and ethical implications within the context of contemporary artistic practices. As explored throughout this essay, the authenticity and intent behind AI-generated artwork are central considerations in evaluating its legitimacy as a form of artistic expression.
The notion of authenticity in art has evolved over time, from the modernist movement’s challenge to traditional notions of authorship and the postmodernist skepticism towards artistic originality, to the current discourse surrounding AI-generated art. While AI lacks inherent intentionality and the capacity for moral judgment, artists like Nils Pooker and Jess MacCormack have demonstrated how human intervention can imbue AI-generated imagery with meaning and authenticity. By actively engaging with AI as a tool rather than a replacement for artistic expression, these artists navigate the complexities of contemporary art-making while preserving their own creative agency.
Moreover, the ethical implications of AI in art extend beyond questions of authorship to issues of copyright infringement and the commodification of creativity. The case of Greg Rutowski highlights the need for clear regulations and ethical guidelines to protect artists’ intellectual property rights in the age of AI-generated content.
Ultimately, the appropriate use of AI in art lies in its integration as a tool for exploration and experimentation within a framework of ethical and aesthetic considerations. As artists continue to push the boundaries of artistic practice, it is imperative to approach AI not as a threat to human creativity, but as a catalyst for innovation and evolution in the ever-changing landscape of contemporary art. By embracing the potential of AI while upholding the principles of authenticity, intentionality, and ethical responsibility, artists can harness its transformative power to enrich artistic discourse and expand the possibilities of creative expression” (OpenAI, 2024).
[1] The technology behind these models is constantly evolving, with recent improvements in video generation.
[2] The modernist movement was catalysed by the industrial revolution, making people eager for a brighter, more developed future under a capitalist market economy. Freedom stood at the forefront of both the American modernist movement, which valued the pursuit of happiness, and the French modernist movement, which highlighted equality and fraternity.
[3] Duchamp signed the urinal as ‘R. Mutt’, in reference to Mott iron works, the manufacturer of the toilet, and Mutt, a popular cartoon character of the time. The ‘R’ stood “for Richard, slang for a rich man” (Foster et al., 2016).
[4] They further challenged notions of authenticity and authorship by not only separating themselves from the making itself but veiling their work in nihilism.
[5] Interestingly, despite using the same reasoning as Duchamp, Koons leans into the idea of the artist ‘as author, visionary and employer’. The notion of the artist being the creative mind, not the maker, is still often critiqued but has been accepted in the contemporary landscape as a viable, sometimes unavoidable, method. The question of authenticity in postmodern art, which has always been veiled in a thick layer of irony, is more difficult to tackle. Jeff Koons has always been controversial in that regard. With pop culture parody and consumerist critiques at the forefront, his highly artificial, shiny sculptures intentionally lack depth of their own. Koons, Warhol and Hirst all made art for the current market, transitioning art into valuable products to be consumed by the higher classes. Art critic Peter Schjedahl said about Koons’ sculptures that they ‘apostrophise […] our present era of plutocratic democracy, sinking scads of money in a gesture of solidarity with lower class taste’ (Foster et al., 2016), touching on their aforementioned lack of depth and capitalistic foundation. Additionally, Koons was ‘found guilty of plagiarism’ (artnet News, 2018) for “stealing advertisement executive Franck Davidovici’s artwork called, ‘Fait d’Hiver’ (1985) [IMG 3], which he transformed into a sculpture of the same name” (Hypebeast, 2018). [IMG 4] This furthermore calls into question the authenticity of Koons’ work, which arguably is an authentic reflection of the postmodern west and the rise of a capitalist free market economy, but leaves the question of personal authenticity and the personal aim of the practice open. The intentional shift away from this defining attitude in recent art history may indicate a push past postmodernism into a post-postmodernist era of art, where both the optimism of early modernists and the nihilism of postmodernists find peace alongside one another.
[6] It is often argued that the cause for this response is evolutionary, as dead or sick humans appear to look not ‘quite right’ and it is in the living, healthy persons’ best interest to avoid it.
“Another explanation holds that the ‘uncanny valley’ results from an inborn fear of death coupled with culturally accepted mechanisms for coping with the inevitability of death. According to this theory, androids evoke our subconscious fears of replacement, reduction, or annihilation. For instance, when androids resemble actual people, they may be construed as doppelgängers.
Consequently, an observer could be afflicted with the fear of being replaced in a certain sphere of life such as in a relationship or on the job. Moreover, androids that are partially disassembled and are depicted in a state of decapitation or mutilation may evoke in the observer pictures of a battlefield in the aftermath of a conflict.
Hence, such scenes can be reminiscent of human mortality. Additionally, the mechanical interior of an almost humanlike robot can evoke the thought that human beings, too, are merely soulless machines.
Furthermore, the mechanical and jerky movements of such an android may elicit the fear of losing control over one’s own body” (MacDorman and Ishiguro, 2006).
[7] In her the book “Robots: Fact, Fiction and Prediction”, Jessica Reichardt later referred to the concept as ‘uncanny valley’.
[8] After Kopernikus, humans no longer considered themselves the centre of the universe and Darwin questioned the idea of us as a mirror image of God. The next big change to the self-concept of humans came with Freud ‘who denied [them] control over [their] psyche’ and finally, “AI is now robbing [the human] of his superior position in the world” (Bahr, 2023). This view on AI is reflected by many figures in the creative world, such as Hayao Miyazaki, the founder of “Studio Ghibli”, who was quoted saying: “I will never apply AI art to my work. The art form is an insult to life itself” (Anime, n.d.).
[9] On his website, he explains that he created a “complete oeuvre of an imaginary US landscape painter […], in the style of a ’missing link‘ between magical realism, abstract expressionism and pop art” (Nils Pooker, n.d.).
[10] Pooker stated that “anyone who feels at home in this style will have to ask themselves what artistic value their own work holds” (Ulrich, 2023), in a pushback against the idea of the artistic genius and desire for a signature style, as well as ”the fetishisation of AI images and their creation” (Ulrich, 2023).
