Photo by petr sidorov on Unsplash
ChatGPT Has Entered the Tarot Reading Room, and It’s Causing Arguments
In A Nutshell
- Researchers interviewed 12 tarot practitioners who use AI tools like ChatGPT to help interpret their readings, finding that self-doubt and the desire for an outside perspective were the biggest motivators.
- Participants used AI to manage uncertainty, generate multiple interpretations of the same spread, and speed up a process that can otherwise take an hour or more.
- Many practitioners worry that leaning on AI could gradually erode their personal intuition, and some expressed skepticism that chatbots are too agreeable to be fully honest.
- When researchers tried to recruit participants through online tarot communities, more than 70% of responses were rejections, with many citing AI as the reason, pointing to a sharp divide within the tarot world.
People have been reading tarot cards for centuries, looking for meaning in a shuffled deck when life feels uncertain or confusing. Now, researchers are noticing a much more modern twist on an old practice. Some pull their cards, then type what they drew into ChatGPT and ask an AI to help make sense of it. A new study takes a hard look at why people are doing this and what it reveals about how humans find meaning, even when the process is designed to be completely random.
The research, presented at the 2026 CHI Conference on Human Factors in Computing Systems, is one of the first academic studies to examine how people use AI tools to interpret tarot readings. What makes this practice especially curious is that, from a scientific standpoint, the cards drawn are not causally connected to the question being asked, even though many practitioners experience the process as deeply meaningful and spiritually significant. So what is really going on?
According to a 2025 Pew study cited in the research, 30% of Americans use astrology, tarot, or fortune tellers at least once a year. That is not a fringe hobby. As AI tools like ChatGPT have become part of everyday life, it was perhaps inevitable they would find their way into the tarot reading room, too.
AI and Tarot: How Researchers Studied the Practice
To understand how people actually blend AI with tarot, the researchers conducted in-depth interviews with 12 practicing tarot readers who already used AI as part of their personal readings. Participants were recruited through Reddit, Discord, Instagram, TikTok, and a local tarot shop, and had to be at least 18 years old, practice tarot or divination, and actively use AI in that practice. Importantly, this was not a survey of all tarot readers. Everyone interviewed had already incorporated AI into their own practice, so the findings reflect how adopters make sense of the tool, not how the broader tarot community feels about it.
The group ranged in age from their late teens to their mid-fifties and came from the United States, Canada, and Mexico. Three considered themselves beginners, six intermediate, and three experts. One participant had been developing personal divination systems for more than 30 years. Interviews lasted about 60 minutes over Zoom; EU and UK residents were excluded due to data privacy regulations. Researchers analyzed all 12 transcripts through multiple rounds of discussion until major themes stabilized.
Why People Turn to AI During a Reading
The most emotionally resonant finding was how many people turned to AI specifically when struggling with self-doubt. Tarot readings can surface uncomfortable feelings or conflicting interpretations, and several participants described a tendency to second-guess their own instincts. One participant, who has spent decades developing personal divination systems, described how AI helped him work through that uncertainty: “I’ve been working on the system for, you know, 30 something years, and I just second guess myself. Am I overthinking this? Am I not thinking enough about this? Am I putting my biases on this too much? … And in the last year, I would say, I really have kind of glommed onto AI as being very useful in helping me sort through the ideas about that.”
Beyond managing self-doubt, many participants used AI to get a fresh perspective, a kind of outside opinion with no personal stake in the outcome. One participant described regularly asking ChatGPT to generate three different interpretations of the same reading, then sitting with each one to see which felt most true: “I will often ask ChatGPT to give me 3 different interpretations. And I will reflect on each one with sort of an open mind and open heart, and think, is this possible that potentially one of these interpretations that isn’t the one I would go to right away actually fits the situation.”
Some participants also used AI simply to save time. A thorough tarot reading can take 30 minutes to an hour, and for people with busy schedules, AI offered a faster route to setting intentions without doing the full interpretive work themselves.
A Tarot Community Divided Over AI
Perhaps the most revealing finding had nothing to do with private practice. When the research team tried to recruit through online tarot communities, they ran into significant resistance. More than three-quarters of their requests to post in forums and online groups never received a response. Of the responses they did get, 70% were rejections, and more than half of those explicitly cited the study’s connection to AI as the reason.
Some moderators had outright bans on AI-related discussions. Participants themselves raised concerns that AI could erode their personal intuition over time. Several also flagged a specific worry: that AI is designed to be agreeable, essentially telling people what they want to hear, which could make it a comforting but not entirely honest interpretive partner.
Thirty percent of Americans already use some form of divination every year, and this study suggests that chatbots are beginning to enter that space, too. Whether AI helps people think more clearly or just reflects back what they already believe remains, for now, an open question.
Paper Notes
Limitations
The study’s authors are transparent about several meaningful constraints on the findings. The sample size of 12 participants is small, and recruitment was genuinely difficult. Because many online tarot communities are hostile to AI, practitioners who use AI in their readings may be reluctant to discuss it publicly, which likely skewed the sample toward people who are more comfortable or open about the practice. The exclusion of participants from the UK and EU due to data privacy regulations also limits how broadly the findings can be applied across different cultural contexts. The researchers describe the study as an “exploratory probe for concept-building” rather than a definitive account. Additionally, the study focused entirely on people who already use AI in their tarot practice, so it does not capture the experiences of those who have tried AI and rejected it, or those who have never considered using it.
Funding and Disclosures
No specific funding sources, grant numbers, or financial disclosures are mentioned in the paper.
Publication Details
Authors: Matthew Kieran Prock, Ziv Epstein, Hope Schroeder, Amy Smith, Cassandra Lee, Vana Goblot, and Farnaz Jahanbakhsh. Matthew Prock and Ziv Epstein are noted as equal contributors. Authors are affiliated with the University of Michigan, MIT Media Lab, Queen Mary University London, and Goldsmiths, University of London. | Journal/Venue: Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI ’26), April 13–17, 2026, Barcelona, Spain. Published by ACM (Association for Computing Machinery). | Paper Title: “Interpretive Cultures: Resonance, randomness, and negotiated meaning for AI-assisted tarot divination” | DOI: https://doi.org/10.1145/3772318.3791571







