Looking forward to being part of
@tarotcultures on Friday 12th June (PM).
John Walter: ‘Stochastic Divination: Tarot, AI and Probabilistic Meaning’ as part of the session ‘AI, RANDOMNESS, AND PROBABILISTIC DIVINATION.’
This paper argues that tarot can be understood not as a mystical system opposed to computation, but as a long-standing cultural technology of probabilistic meaning-making. Drawing on practice-based research developed through ‘The Tarot Garden,’ ‘Alien Sex Club Tarot,’ and ‘Lockdown Tarot,’the paper situates tarot as an interpretive system structured around recombination. Across these projects, tarot decks were conceived as mutable image systems, produced through vector and VR drawing, installation, and tested through live readings.
The paper extends this argument through analysis of ‘Stochastic Tarot,’ a deck developed in dialogue with the concept of the “stochastic parrot”. Research for this project involved experiments with generative AI to explore stylistic variation and image mutation, particularly in the generation of borders and ornamental structures. The final deck, however, was drawn by the artist in Adobe Animate as a complete 78-card system, maintaining authorship while embedding stochastic logic within the imagery.
Subsequent stages of the research reintroduce AI as an interpretive rather than generative agent. During algorithmic tarot readings, selected cards from the ‘Stochastic Tarot’ are fed into openart.ai to generate verbal–visual descriptions of what the machine “sees”. These descriptions are then used to produce unique image variants of the selected card, functioning as on-demand artefacts for the querent.
Through case studies drawn from live readings in which these variants are generated, interpreted and printed in real time, the paper examines how distinctions between tarot card, prompt, reader and machine intelligence begin to blur.