AI psychics are already a business, and their opacity over intimate data is the real problem, not superstition

🕒 Published on Zendoric: October 30, 2026 · 00:05
✨ AI-generated · how it's made
Apps like Your AI Psychic sell readings for coins while human psychics add "no AI" clauses. Behind the anecdote lies a serious question: who protects the confidences people share with a chatbot.
By The Guardian · October 29, 2026.
Automated fortune-telling is now a consumer product. According to Kari Paul's report in The Guardian, the app Your AI Psychic offers more than a dozen bots that give readings via chat, video and voice, as well as palm reading and astrological charts. It runs on coins: each question costs 4.06, and packages range from $9.99 (82 coins) to $99.99 (1,180). AI psychics are also appearing on TikTok Live, and according to human practitioners, Etsy is filling up with AI-generated horoscopes and spells.
Human professionals are responding defensively. Lauryn Kellett, a psychic with ten years in the trade, says she teaches people to tell the human from the artificial and encourages her clients to post videos and photos of themselves to prove they are not bots. Others add "no AI" clauses to their listings. Ahna Hendrix, who works with the so-called Akashic records, argues that relying on AI to validate decisions weakens each person's inner guidance.
The strongest part of the report is not about the supernatural but about transparency. The main app does not mention AI in its privacy policy, referring only to "psychic advisors," and does not name any model provider. According to the outlet, neither the app's publisher (RF Master Group Ltd) nor its purported parent company ("Expert Center") responded to requests for comment. Another platform, MystiCall, did respond: it says it uses ElevenLabs for voice and OpenAI for summaries, that it does not sell conversations or use them for targeted advertising, and that it has disabled training on data with its partners. It acknowledges that its policy allows conversations about mental health, sexuality or finances. These are the company's own claims, not verified by us.
Anthropologist Sally Applin provides the most useful framework. With Michael D. Fischer, she argues that consulting an LLM (a language model that predicts text based on statistical patterns) is a modern version of consulting an oracle: the meaning is supplied by whoever interprets the answer, and the question is repeated until the answer is pleasing. The key point is that a "psychic" app isn't needed for this: many people already ask ChatGPT whether to change jobs or leave their partner.
**Our take.** Mocking the AI psychic is easy and uninteresting. What matters is that these apps concentrate the most sensitive parts of a human conversation (fears, breakups, money, health) in an unsupervised product, with opaque models and data policies that don't even acknowledge that an AI is involved. The risk here is not whether the bot gets it right or wrong: it is that many users entrust it with intimate matters without knowing where they end up. The report also cites studies suggesting that LLMs can amplify paranoia in people vulnerable to psychosis; a people-pleasing chatbot that says "you'll get married soon" is harmless, but one that validates a harmful belief is not.
We should also be fair to the other side. Laurie-Anne Bernard, co-founder of MystiCall, argues that AI removes cost and availability barriers, and it is true that such access matters for those who cannot afford a professional. But that logic is exactly the same as in the debate over AI therapy, and there, what separates a useful service from a harmful one is clear limits: no predicting, no diagnosing, no recommending high-stakes decisions, and disclosing who processes the data. The human psychics who refuse to weigh in on health or finances on TikTok Live are applying a safety rule that many AI apps lack.
In the short term, expect more noise: synthetic content passing itself off as human, and "made by people" certifications as a signal of value. More broadly, when the cost of producing a service falls to nearly zero, what becomes scarce and commands a price is verifiable authenticity and trust. In the long term, the same kind of well-governed assistants, with protected data and honest limits, could offer accessible emotional support to people who today have nothing. Whether they become that, or merely a coin-based business built on sensitive data, depends on the regulation and transparency demanded starting now.
🔗 Related on Zendoric


