The Great Clash over Creation

Earlier this year, a new viral trend took social media by storm yet again. OpenAI, the same company that developed the generative artificial intelligence (AI) assistant ChatGPT, launched a new feature that allowed users to upload and alter photographs to resemble the hand-drawn animation style of Japanese artist and filmmaker Hayao Miyazaki. Miyazaki’s name is synonymous with Studio Ghibli, the animation studio he co-founded to produce award-winning films such as Spirited Away, Kiki’s Delivery Service, Howl’s Moving Castle, and My Neighbour Totoro which have enthralled viewers with their heart-warming stories and incredible visuals. In a flash, Instagram was inundated with Ghibli reimaginations of user’s personal photos, scenes from iconic films, celebrities, and natural attractions such as Darjeeling and Munnar. The feature generated a divide between those who jumped in on the opportunity to reimagine themselves as ‘Ghiblified’ avatars, and others who accused OpenAI of plagiarism, copyright infringement, misappropriation of an artist’s style, and privacy violations. As pertinent as these concerns are, the tug-of-war between OpenAI and Studio Ghibli points to an issue far greater than the legalities of intellectual property. The surge in AI tools that promise to mimic and even outperform human cognition, creativity, perception and emotion points to a looming crisis in our human condition.

It is important to reiterate at this juncture that art and AI are not rivals. In a sense, AI is only the latest in a long line of technologies developed to expand the possibilities of an artist’s creative output. The scepticism to which AI is currently subject was once levelled at digital art tools such as Procreate, computer animation, and digital photography—to wit, these tools automate so much technical skill that their use amounts to cheating, and their immaterial nature takes away the immediacy that we associate with traditional art forms such as oil paintings and sketches. That these art tools have now become standard in both commercial and fine art is testament to the weakness of these apprehensions. The widespread adoption of these tools has not only encouraged the development of new skill sets and artistry; the works thus produced have also generated their own unique forms of tactile immediacy, notwithstanding their virtual nature. So far as AI is concerned, both academic literature and artists’ reviews have testified to the ability of AI tools to enhance both the creative output and the artist’s experience of creating it. A report by the Oxford Internet Institute, University of Oxford, found that art made through machine learning opened up new spaces of artistic agency, multiplied sources of creative inspiration, reorganised creative workflows, and opened up new technical affordances whereby artists could exploit not only the abilities of machine learning but also its glitches to produce interesting variations in their work. Most importantly, the artists maintained that the adoption of machine learning was just another change in medium and tool, not a siphoning off of their own creative intent.

However, the passive, one-prompt-away AI art of the kind OpenAI sought to popularise should ring alarm bells for the simple reason that it foreshadows our collective deprivation of one of the faculties that makes us fundamentally, irredeemably human: the ability to make, intentionally. Philosophers ancient and, well, less ancient have agreed that one of the ways in which humans stand apart from other animals is in our capacity as homo faber, or as makers. Making is not confined to craftsmanship alone, but also encapsulates the ability to plan, to practice, and to make and learn from mistakes. The anthropologist Tim Ingold urges us to reconsider making not as the simple executive application of a series of operational steps, but as a generative process that opens up opportunities for reflection, growth, and the occasional creative spark. As a knitter, I can confirm that the joy of knitting lies as much in the interstices of stitches—where I learn to gauge the tension appropriate to different kinds of wool, where the pattern suddenly makes sense to my muscle memory, and where the movement of strands between the needles sparks inspiration for a new project—as it does in beholding the finished product. The process of making anything—clothes, food, or even a YouTube video - offers time and space to do things deliberately. It prompts us to slow down, check in with ourselves, and grounds us in the here and now.

What also distinguishes making as a uniquely human phenomenon is the collaboration it involves. In her seminal work The Human Condition, German-American philosopher Hannah Arendt describes what she refers to as the vita activa, or a life engaged in collective action, as providing both the physical and cultural durability required for human life to progressively unfold. Where art is concerned, collaboration entails the sharing of knowledge, practices, and networks accumulated over time and across space, but it also involves affective exchange—the artist’s experience of their relations with the subject being depicted, and the resulting intention. Consider, for instance, a pair of paintings made by the 17th-century Dutch painter Frans Hals of the Governors and Governesses of a Dutch alms house. English art critic John Berger described the drama of the paintings, where Hals chose to consign a bitterness of expression and a drunkenness of manners to the men and women who sat before him. According to Berger, this artistic choice sprang not from some fact of subjective observation or thematic decision on Hals’ part, but rather from the conflict that the destitute painter faced in having to make officially commissioned portraits of the very people who oversaw his destitution. Such social tension can hardly be expected from an application that mistakenly adds extra limbs to images of users, let alone an element of conflict inspired by its own lived reality. Why, then, are we keen to outsource a faculty so central to human experience to unfeeling machines that masquerade their adaptive algorithms as emotion, thought, and personality? Why is it that an AI programme which mimics human sentience and intention sparks more popular excitement than ones which, for example, can potentially eliminate the discriminatory practice of manual scavenging? The unfortunate fact is that we live in a digital ecosystem that feeds on instant gratification, where the patience, effort, repetition and, often, failure that underpin any human enterprise are made redundant by instantly available ready-mades. We do not make things as much as we accumulate them. We do not develop opinions as much as we consume them. The instant validation from parasocial followers proves stiff competition to the depth of emotional exchange and the risk of heartbreak required to cultivate meaningful relationships. Such an ecosystem neutralises our propensity for a life of collective action by reducing us to passive onlookers to our own lives. For inactive individuals isolated from each other, the humanoid functions of easily accessible technology become a welcome substitute for the vita activa.

The technocrats do not care—they rarely do. In the wake of the controversy with Studio Ghibli, OpenAI CEO Sam Altman’s overriding concern was that the organisation’s GPUs were overheating due to the demand for Ghibli-style images. The decision thereafter to impose rate limits on the use of ChatGPT to generate such images was strictly a business move. At the time of writing this article, the ChatGPT feature continues to be available to produce images in various art styles, including that of Studio Ghibli. A host of other AI platforms now offer similar services. Between the protestations of paintbrush-wielding purists and Altman passing off appropriation of an art style as a harmless case of ‘art inspiring art’, there is surely a threshold where one crosses over from using AI to make art to mindlessly generating images. While debates over legal redressals to the issues raised by various stakeholders continue to rage, it may be worthwhile to ask ourselves what it is that we risk losing when every outcome is made available to us just a prompt away.

The writer is a doctoral researcher in anthropology at the University of Oxford.
Views expressed are personal