When I pull a traditional Major Arcana card, there's centuries of symbolic density behind it. The Fool, The Tower, The Lovers—these aren't just illustrations; they're psychological anchors. Jung called them archetypes, universal patterns that resonate across cultures. AI, on the other hand, is just predicting the next token based on training data. It doesn't "see" the symbolism; it statistically approximates it.

I've noticed a strong confirmation bias at play. Because I'm navigating PTSD and using meditation to ground myself, I'm naturally looking for the cards to reflect my inner state. But AI oracles tend to give you exactly what you already expect. It's like looking into a funhouse mirror. The algorithm optimizes for agreement, not revelation. Real divination works through projection—you stare at a fixed symbol and let your subconscious fill in the gaps. AI fills in the gaps for you, which actually short-circuits the psychological process.

That said, I'm not dismissing them entirely. I've seen how structured prompts can produce surprisingly coherent spreads, and for casual reflection, they might work. But if you're looking for genuine psychological depth, you still need the static, unchanging symbols that have survived centuries. The machine can generate the image, but it can't generate the meaning. That part still has to come from you.

Has anyone else tested AI oracles alongside traditional decks? I'm curious if the lack of symbolic consistency ruins the spread, or if you've found a workaround.