AI oracles are all the rage on TikTok. | John Lund/Getty Images
TikTok users are turning AI filters into fortunetellers.
Can an AI predict your fate? Can it read your life and draw trenchant conclusions about who you are?
Hordes of people on TikTok and Snapchat seem to think so. They’ve started using AI filters as fortunetellers and fate predictors, divining everything from the age of their crush to whether their marriage is meant to last.
The most viral instances have featured broken-hearted women waxing despondent because their AI filter is erasing the men in their lives from photos of the two of them. To the women, this either means they will never find love again, or that the love they had was doomed.
The trend of using AI signifiers as oracles has taken off more generally as well. TikTok is awash in people swearing fealty to everything from randomized “soulmate” predictors to third-eye detectors. The “soulmate filter” tag has racked up over 30 million views for a bevy of options such as filters showing how far away your soulmate is, when they were born, how compatible you are, and whether you have an AI-generated soulmate ring (indicating you’ve already found your soulmate). Others are using AIs to build on tarot readings, see chakras, or generate horoscopes that don’t sound quite like the ones in your local newspaper (“the stars recommend being in a state of quantum antelope”).
For the most part, these memes aren’t serious, and most come across as pure silliness — though occasionally, things get a little more alarming. One purported actual widow posted her AI-generated oracle erasing her dead husband’s photo. AI may also start detecting ghosts all around your house.
All of these glitches in the matrix may stand out as part of a viral TikTok trend, but they also speak to a larger cultural desire for artificial intelligence to be more than what it is. In its current stage of development, AI is nothing more than a giant collection of data points that can be shaped into predictive patterns. There’s nothing sentient or supernatural about it.
Yet humans inevitably seek ways to humanize AI — and, really, this shouldn’t come as a surprise. Humans read into everything. We anthropomorphize robots. We pretend our pets love us. We imagine our lives are being directed by everything from sky gods to playing cards to star patterns.
“It’s fundamentally human for people to want to play with and explore these technologies,” says Karen Gregory, a sociology professor at the University of Edinburgh. “This is the essence and history of divination and gambling. Cards, bones, tea leaves, all manner of objects (whether they are digital or not) can be used to play with change and uncertainty — to play with the question of ‘what next?’”
She explains that “at some level, we are compelled to play.” So it seems inevitable that humans are doing their best to make auto-generated AI bots into predictors of fate.
“In general, we tend to scan for patterns in our environment,” A.J. Marsden, an associate professor of psychology at Beacon College in Leesburg, Florida, tells me. “Our brains are trying to make sense of the world around us and search for any potential threats. Finding patterns helps our brain process information faster in order to make predictions regarding future events.” In other words, with these virtual oracles, we’re doing what we’ve always done — just with newer, quirkier tools.
“What’s happening here is akin to getting ‘a reading,’ much like you would get from a psychic reading or a tarot card reading, except here the subject is your media content,” Gregory tells me. “I think it makes a lot of sense that people are playing and experimenting with TikTok’s AI filters. Social media norms and the broader social context of creating content encourages and rewards this kind of experimentation and meme making.”
Gregory points out that the internet has already given rise to countless forms of digital innovations in New Age practices. Take, for example, the wildly popular Labyrinthos tarot, which comes with a fancy mobile app that auto-generates card readings and meditations with the click of a button. She also notes that there’s a striking similarity between what people are doing with AI online and what people already do with tarot cards — perhaps because the mechanisms we use for making meaning from random results are virtually identical.
“Like a tarot card that has been flipped, whatever the AI generates can be read as personally meaningful and significant, says Gregory. “In my own work on tarot communities, I’ve looked at people’s relationships with tarot cards, and what is happening here seems very similar — the power of the card flip to quickly and almost effortlessly produce something new, insightful, and useful is being found in the AI filter’s response. Once a card has been flipped, your attention has been pulled into a next moment in time, a next possible interpretation. That’s exceptionally valuable in a highly uncertain world.”
You might think from this type of observation that AI fortunetelling would fit seamlessly into established esoteric traditions. Yet while many people are optimistic about the creative possibilities of tarot, others hotly debate their use in the creative sphere of divination.
Some insist that the spiritual essence of divination gets lost when you try to map the creative process onto a machine. In April, longtime druid and tarotologist Dana O’Driscoll wrote a lengthy blog post arguing against the use of AI in divination for a number of reasons. Many of O’Driscoll’s arguments are familiar to anyone who’s followed the ongoing debate over AI and creativity: AI is no substitute for artistic inspiration. But O’Driscoll went further, expressing concern that a reliance on AI pulls people away from a connection with all of the inner spiritual insight that divination is meant to cultivate. “The broader problem as I see it,” she wrote, “is that in mechanizing the world and in turning people into consumers, we’ve also seen a major loss of a really important thing for human development and consciousness — the cultivation of a rich inner life and a deep connection to nature.”
Since O’Driscoll’s animistic philosophy holds that all things, even artificial intelligence, are imbued with a spirit, her concerns enfold the worry that “since AI has been created for obvious capitalist reasons,” the spirits of the machines might be of dubious intent. With AI tarot decks and other metaphysical tools, whatever spiritual energy might be present fills her with skepticism. “What I say is that under no circumstances will I touch anything spiritual that has been created with AI,” she wrote. “Tread very carefully, friends.”
Of course, from a skeptical viewpoint, there are plenty of non-metaphysical reasons that AI-generated oracles are bad news. “Although fun to engage with, in almost every case, divinatory tools are merely coincidences,” Marsden reminds us.
Marsden does concede that many people seem to find a psychological benefit in using tarot and other divination tools as forms of self-reflection. “If I use tarot cards as a way to predict the likelihood of finding a partner, there probably won’t be much benefit psychologically. If, however, I use tarot cards as a form of self-reflection — what am I looking for in a partner, what would make me happy, etc. — then the cards would likely have more benefits to us psychologically.” Science suggests, she notes, that when we do things with intent, like spell-casting or goal visualization, we’re more likely to work hard to achieve the things we want.
Ultimately, though, she argues that any psychological benefit derived from such esoteric tools might not outweigh the cost of deluding ourselves into believing divination is real. “We often find patterns in random phenomena, so any shortcuts based on these patterns would not be reliable. And therein lies the problem with divination. In most cases, we are likely deluding ourselves more than we are truly benefiting ourselves.”
Still, despite being blunt — “AI does not have magic powers, it does not know you better than you know yourself, and it is not revealing special information that should be banked on or trusted” — Gregory holds that it’s a very human trait to be drawn to the promise of an AI oracle. “Much like [a psychic’s] cold reading, the AI doesn’t have any special powers. It’s responding to input and spitting out a response. However, whatever that response is — something beautiful or even gibberish — it can become the grounds for new meaning or new interpretations.” She describes the phenomenon as “a great extension of a very human curiosity to see what comes next and to manage that uncertainty and anxiety.”
And if much of that uncertainty and anxiety involves the question of AI itself, well, perhaps the humanization of your TikTok filter may be all the better to welcome our robot overlords.