How far can the resonances of art history, shared cultural knowledge, and language echo in the
age of Artificial Intelligence (AI) and machine learning? Can such rapidly developing technologies play a meaningful part in the broader art historical schema? These are some of the questions that AI (Art Index), the new book from Vienna-based Australian artist James Tunks, broaches in its dialogue of historical texts, images, and their machine-generated counterparts.

Exploring the format of an art history textbook, the project uses the language of artworks as its starting point, combining original works from museum collections with images created by cutting up and fragmenting various written resources from museum collections – artwork descriptions, keywords, and exhibition texts – before feeding them into machine learning text-to-image generators. The results prove provocative and perplexing, teetering between improbable likeness and alien flourish.

The application of techniques such as machine learning and AI image-making tools not only raises the question of whether machines can truly think, but also whether they have the capacity to interpret and visualise abstract ideas and concepts synonymous with the history of art. In this re-reading of some of art’s canonical signposts and footnotes, AI (Art Index) helps open a new line of communication.