Zsolt Almási: AI? Predictive Media. Art?

In this lecture, I shall delve into the relationship between predictive text2image technology and art. Text2image applications have been with us for decades, but they have become famous (maybe infamous) since Open AI released Dall-E for the public in the summer of 2023, closely followed by akin platforms such as Stable Diffusion, Midjourney, and Nightcafé Studio. I will discern some early responses from photographers to this technological innovation, drawing analogies to the advent of digital photography. This is done so that a pivotal question may emerge—can this nascent media technology be aptly denoted as art? To contemplate this, I shall navigate the discursive landscape that facilitates the consideration of this cultural phenomenon within a rational framework. This is necessary, as public discourse is mostly dominated by unrealistic expectations resulting from identifying this technology as “Artificial Intelligence.”

The theoretical frame is posthumanism, affording us the necessary intellectual tools to scrutinize the genesis of what Manovich calls “predictive media.” Equipped with the appropriate terminology, I will explore the process of the creation of images with a special focus on prompt engineering and curation to showcase the collaborative effort of the artist and the application. Furthermore, we shall refer to the market (Christie’s) and the competitions (Sony, Főfotó) wherein this technology plays a consequential role, and to traditional art forms that share some crucial features with this technology, furthermore some concerns related to predictive media. Through this analysis, I shall argue that contrary to early responses, images generated with text2image applications bear the hallmark of art, and their creators rightfully merit the designation of artists.