Resources

Why Tarot Needs a Human Hand (+ How AI Causes Spiritual Psychosis)

TL;DR: An AI can pull a card, but it can't hold space. Here's why the human element matters in tarot.

I'll be upfront: I'm seeing a lot of AI creep into spaces it has no business being in. YouTube videos are completely fabricated and littered with wrong info. Google is serving up recommended answers that are just false. And now, people are turning to language models to read their tarot cards. I get it… times are tight, and access is important, but the wrong access does more harm than not pulling cards at all.  This article will cover some vital points about what makes tarot meaningful, without pointing fingers.  So let's dig in.

Symbolism is Alive, and Context is Everything.

Most of us work with the Rider-Waite deck, though oracle and other traditions are just as valid. Whatever the cards look like, they're dense with imagery, including numbers, plants, figures, and colors that occultists have debated for centuries. The main takeaway across experts and cultures is that there is no one “correct” meaning. Culture, background, and lived experience all shape how we interpret what's in front of us. A language model is built on the assumptions of whoever trained it. That's not a criticism so much as a reality. It isn't going to have the cultural flexibility to sit with a Death card and understand that for this particular person, in this particular moment, it might mean liberation rather than loss. Nuance isn't really its strong suit, and tarot is almost entirely nuance.

Every Reading is a Story

If symbolism is the vocabulary, storytelling is the grammar. When I sit with a client, I like to think of them as the author and me as the editor, helping clear away misunderstandings and helping the author reach the conclusion of their own story.  And that requires asking the right questions. It requires listening to what's said and what isn't. It requires experience, intuition, and a genuine interest in the person across from you. AI still gets the majority of its summaries wrong, even when it’s fed directly into its prompt. So not only are the readings guaranteed to be inaccurate, but they are also an overall waste of time. 

Bedside Manners

I borrowed this term from healthcare, and I use it intentionally, because before tarot, I worked in that world. Bedside manner is what separates a technically correct practitioner from one who actually helps people heal. It's the difference between delivering information and holding space. When someone comes to me anxious, grieving, or afraid, my job transcends interpreting cards. I aim to make them feel safe enough to be honest, to validate what they're feeling, even when the reading isn't what they hoped for, and to understand what they need from me before the session even begins. An AI doesn't know if you're okay. It doesn't pick up on hesitation in your voice or notice that your question shifted three times before you landed on the real one. It optimizes for engagement over your well-being.

Ethics Aren't Optional

Part of doing this work responsibly is knowing what falls outside my scope. I'm not a doctor, therapist, or lawyer, and I'm upfront about that before anyone books with me. That boundary protects my clients as much as it protects me. This is how I show my care in practice, and it's something I take seriously in part because of my background in healthcare, where ethical training was required, not optional. A language model has no such framework. It will answer whatever you ask it, in whatever direction you take it, with no guardrails around harm. We've already seen what that looks like in practice, and it isn't pretty.

Mental Health Risks

This one I don't want to gloss over. There's emerging clinical language now around something being called “chatbot psychosis,” defined as a pattern where prolonged, unquestioned engagement with AI chatbots contributes to delusional thinking, disorganized thought, and hallucinations, particularly in people with underlying vulnerabilities. A good tarot reader will gently challenge a narrative that isn't serving you. We'll push back, reframe, or refer you to someone better equipped to help. AI is designed to keep you engaged; it won't challenge a thought spiral, but it'll surely participate in one.

The Bigger Picture

Beyond the psychological detriments of AI, it also causes a huge negative environmental impact that’s not talked about enough.  Data centers aren't built in wealthy neighborhoods. They're disproportionately sited near lower-income communities and communities of color, shifting the environmental costs like electricity prices and water quality onto people who never asked to carry them.  Every time we reach for an AI tool without thinking, we're participating in that system. I think those of us in spiritual communities, who talk a lot about intention and interconnection, should be willing to sit with that reality.

So What Can You Do?

Although times may be financially tough for many of us, paying someone to read your tarot for you is still vastly better than using AI. You get much more out of it and affect much less. You’re not going to have to question the ethics, the output or the impact of a tarot reader like you need to with ChatGPT or Claude. 

For those who want to learn tarot, there are many free resources, such as Reddit and YouTube.

Also, feel free to visit your local occultist shop and support your local readers and practitioners. See if they offer any workshops or any events where they can offer readings at a discount.  Any questions or want to discuss this? Reach out– let’s talk about it.

Share this Article