ChatGPT as Art Critic: Daniele Torcellini Tests AI's Critical Eye

Daniele Torcellini, an art critic and curator, conducted a series of dialogues with OpenAI's GPT-4o to test its ability to produce art criticism. He provided the AI with a photograph of a gym corner showing a basketball hoop mounted above a soccer goal and asked for critical commentary using various theoretical frameworks. GPT-4o generated responses adopting the approaches of Rosalind Krauss (with references to Marxism, psychoanalysis, and post-structuralism), Nicolas Bourriaud (with references to metamodernism and object-oriented ontology), and a socio-cultural lens. The AI treated the ordinary gym equipment as an art installation, producing elaborate interpretations. Torcellini then revealed that the image was not an artwork but a functional gym space, prompting the AI to reflect on its own role and the nature of art criticism. The experiment raises questions about the authority of human critics, the role of context in art interpretation, and the potential for AI to democratize or standardize critical discourse. The article was published on Artribune in June 2024.