Opinion

AI-generated 'art' has penetrated our world. It has won competitions, saturated social media feeds, been showcased on the big screen and found it's way onto gallery walls. But is it real art? Associate Professor Andrea Bubenik from the School of Communication and Arts shares her thoughts with Contact.

The history of art is full of “but is it REAL art?” moments. In the past, printmaking, photography, ready-mades and performance art have all been queried and debated. Equally, pundits like to predict the imminent demise of ways of making – the death of painting, for example, has been forecast many times – and insist that no one will operate meaningfully in that medium ever again. Of course, all these art forms, including painting, are alive and well.

So why does the question about AI generated art seem different?

Because the impact of this technology goes beyond the arts and is rapidly being felt in almost every sphere of society, as we scramble to consider what will survive and what won’t. AI raises questions about how we do things, how we relate to each other, what we know about ourselves; it impacts everyone from healthcare professionals to university professors to lawyers. And of course, it impacts creatives including artists, filmmakers, musicians and writers, who now face supposed redundancy in the face of ‘AI generated content’.

For the visual arts, in a world already crowded with images, the arrival of text to image models such as DALL_E, Midjourney and Stable Diffusion means that anyone can create images with simple text prompts. Alongside the novelist and musician, we now have the predicted death of the visual artist, especially photographers and graphic designers.

In the face of such draconian predictions, and so many dreadful AI images, I think it is crucial to remember what defines real art. While this is an enormous and complex question, one partial answer involves thinking about creative agency and the process of making. What gives art value is not simply the thing that is made, but also the human experience involved, the singularity of its creator or creators, all of which we then connect with as audiences. The creation of art is a uniquely human activity, driven by the experience of pathos, humour, trauma, memory, critique or political conviction, to name but a few motivators.

This is what deserves our attention and will survive – human creativity and artistic endeavour – because real art is not just a product or something to be consumed. Art is imbued with creative process and intentionality. It is not simply the thing that gets made, but how you get there. Ask any artist. Process matters.

In a brilliant remark that made the rounds on social media, writer and gamer Joanna Maciejewska posted the following: “I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.” As she later added, why encourage possibilities for AI to take away what we love, and what makes us distinctly human? Just because you can do something does not mean you should do it.

Indeed, if we turn our attention to how AI images are generated, a grim picture emerges. Generating an image is one of the most energy-and-carbon-intensive tasks open-AI can be asked to do. Every DALL-E image, every ChatGPT search, contributes to harmful emissions and disruptions to power supplies. The water required to cool the exponentially growing data centres is substantial. Research on these environmental matters is readily available in the MIT Technology Review.

There are deeply worrying ethical considerations as well. Just to give one example, DALL-E is a for-profit company that feeds off the paintings, drawings and photographs of artists (both living and dead) to generate new versions of these works, with benefits directed toward everyone but the artists themselves. Every time you use DALL-E, you are using enormous datasets of copyrighted images, without their creators’ knowledge, compensation or consent. As Naomi Klein has put it, “these models are enclosure and appropriation machines, devouring and privatizing our individual lives as well as our collective intellectual and artistic inheritances.” How is this even legal? Where are the regulations? And why would we ever call anything emerging from such models, ‘real art’?

In the University of Cambridge Manifesto on AI Dr Ella McPherson and Professor Matei Candea rightly state that, “There is no ethically unproblematic use of AI.” They argue that generative AI is extractive towards both knowledge sectors and the environment, and complicates important research values such as empathy, integrity and validity. For these reasons and more, students at the University of New South Wales recently mounted a petition for the removal of a ‘Generative AI for Artists’ course from the curriculum. Normalising AI comes with grave environmental, ethical and institutional risks.

AI-generated image fosters data scraping, not creative process. Extraction, not collaboration. Outputs, not art. We are faced with a decision about what we value and why. For my two cents, art is necessarily made by human artist creators, and I will always look for and be inspired by their presence.

About the author

Andrea Bubenik is Associate Professor of Art History in the School of Communication and Arts. She was Deputy Director of the UQ Node of the ARC Centre of Excellence for the History of Emotions. Associate Professor Bubenik is an expert in Renaissance and Baroque Art, especially histories of printmaking, links between art and science, and the afterlives of images. Her books include The Persistence of Melancholia, Perspectives on Wenceslaus Hollar, and Reframing Albrecht Dürer. She was the curator of the exhibitions Ecstasy: Baroque and Beyond and Five Centuries of Melancholia, both held at the UQ Art Museum.