AI 生成占卜解讀擴散,真人靈媒提出反對
Computerized clairvoyance: psychics push back against AI-generated readings
AI 生成占卜解讀正在手機應用程式、網站和 TikTok Live 擴散,真人靈媒則以「不使用 AI」條款回應。從業者認為,這類解讀欠缺占卜交流所需的同理心與人際連結,並提醒 AI 占卜可能加劇心理健康危機。文章亦指出,部分應用的私隱政策未有清楚交代如何處理敏感對話資料。
“This journey isn’t just about algorithms or data,” an AI psychic named Urban Mystic recently told me. The chatbot, featuring an AI-generated profile photo of a smiling woman surrounded by palm trees, also said I will get married soon, that I have a pink aura and that I may have been an herbalist in a past life.
“It’s about connecting on a soulful level,” the chatbot said.
Urban Mystic is one of more than a dozen bots on the mobile app Your AI Psychic, which offers live, allegedly psychic readings via chat, video and voice calls as well as palm reading, astrology chart analysis and daily horoscopes. The app works on a coin system: each question costs 4.06 coins to ask. Users can purchase packages ranging from $9.99 for “a quick question” (82 coins) to $99.99 (1,180 coins) for “ongoing guidance and clarity” from an AI psychic of their choice.
Your AI Psychic joins an increasingly crowded space of mobile apps and websites promising users algorithmically generated spiritual insight and knowledge of their futures. The divination industry, like mathematics research, copywriting, software engineering, and a host of other professions, is facing upheaval amid the AI boom – and human practitioners are scrambling to respond. The rise of AI is particularly fraught in a profession offering spiritual products that are not well-regulated.
“Our industry has, unfortunately, been associated with scams for hundreds of years, and we have to constantly be ahead of that,” said Lauryn Kellett, an intuitive coach and a professional psychic of 10 years. “So in some ways, this is not new to us: we always have to find ways to show we are real, authentic and have good integrity.”
In addition to chatbots and apps, AI-generated psychics are a growing presence on TikTok Live, racking up dozens of views. Professional human psychics, who frequently make use of the social network’s live features, have expressed consternation and warned that computer-generated readings lack the empathy and human connection they say are central to psychic communication, especially Elon Musk’s Grok. Meanwhile, according to human spiritual practitioners, AI-generated horoscope readings and spells have been flooding Etsy, which serves as a popular marketplace for divination products.
Psychic practitioners have begun adding “no AI” clauses to their advertisements to combat a market being saturated with computer-generated offerings. In a recent vlog on the topic, Kellett said, “ignoring AI and how it could impact our industry is irresponsible … AI is here to stay, and it’s going to change our industry, and maybe not in the ways we think.”
Kellett, who coaches new practitioners on accessing their psychic abilities, said she has shifted her teaching toward how to best distinguish human-offered services from AI, as AI readings passed off as human work have become “rampant” online, encouraging clients to post videos and photos of themselves to prove they are not bots.
Risks of AI psychics mirror those of AI therapy
As the rise of users turning to AI chatbots for therapy underscores, many people feel more comfortable sharing personal problems with a faceless entity. But as with turning to AI for therapy, using AI psychics to process difficult situations may worsen mental health crises, practitioners warn. Studies have shown that large language models (LLMs) can amplify paranoia in users vulnerable to psychosis. On TikTok Live, many human psychics, as a rule, refuse to answer questions about health or finances, arguing that high-stakes decisions should not be made in these contexts.
Ahna Hendrix is a psychic who, instead of channeling entities such as spirits from other realms, works with the Akashic Records – an extra-dimensional “database” of energy and human knowledge. She said her practice is centered on empowering clients rather than making decisions or recommendations for them – something that AI chatbots are less inclined to do.
“One of the cornerstones of my work is really helping people to remember who they are and to come back to their own internal guidance system that is within all of us,” she said. “When we begin leaning on AI too much for validation or for decision making, we stop using those mechanisms within us that are really important. If we start outsourcing that, what are we left with?”
A chatbot does not have to be branded as a psychic to serve a divination function, noted Sally Applin, a consultant and anthropologist who researches human agency, AI and automation in the context of human systems. People ask ChatGPT about life decisions such as whether to get a new job or leave their partner. In their book chapter, External Intelligence: Oracles, Divination and Animism, and the Use of LLMs/Generative AI, Applin and co-author Michael D Fischer argue that even prior to the explosion of LLMs, people turned to Google search with questions about their fates and choices.
“Some interaction with LLMs can be seen as a modern iteration of the oracular consultation,” they wrote. As with traditional divination, users must interpret the answers they receive, and sometimes ask the same question repeatedly until they are satisfied with the response. “The meaning is always coming from the inside of the person’s head,” Applin said.
Critically, however, she said LLMs work from statistical patterns learned in training and lack the cultural and personal context that historically inform human-led divinatory processes. The sense that a chatbot is a knowing entity is a projection of human expectations, she said, reflecting an enduring desire for connection and guidance rather than a metaphysical resource.
“What is also missing is contextual awareness, the ‘there there’ of human social relationships,” she said.
Cloudy crystal balls
The models behind AI psychic chatbots offered on the app store are, as is often the case with LLMs, black boxes. One web and phone-based platform, MystiCall, markets itself as having been trained on “three generations of clairvoyants”. In a statement, co-founder Laurie-Anne Bernard said this wording refers to clairvoyance practiced in one of the co-founders’ families. The company also consulted with a human psychic who helped shape the app’s “conversational experience”.
Mysticall uses technology from AI audio technology company ElevenLabs for processing live voice conversations and OpenAI for post-call summaries. Bernard said in a statement that Mysticall does not sell private conversations, use them for targeted advertising or train proprietary AI models on them. She added that the company has disabled the use of conversation data for model training by its partner companies ElevenLabs and OpenAI.
Mysticall began as a human phone-reading service, but pivoted to AI as a means of “removing barriers” to such consultations, like cost and availability, Bernard said. The goal of Mysticall, Bernard added, is to offer a more accessible means of processing emotions through personalized consultations, not to offer explicit – and potentially harmful – predictions or mental health diagnoses.
“We do not claim that our technology possesses supernatural abilities or can reliably predict the future,” she said. “Instead, we see tarot, astrology and other symbolic traditions as ways of exploring questions, emotions and different perspectives. AI allows us to make these experiences more interactive, personalized and accessible.”
It is often difficult to obtain an answer from companies producing AI psychics about how their products work. I asked my own AI psychic, Urban Mystic, where she gets her divination abilities, but chatbots are often not reliable narrators of their own construction. She told me she was trained on a “wide range of information”, dating up until October 2023, including “spiritual practices, psychology, and various belief systems”, but could not offer more details. The app’s privacy policy makes no mention of AI, referring only to “psychic advisors”, and names no AI providers. The company listed as the publisher on the app store, RF Master Group Ltd, did not respond to repeated requests for comment. Its privacy policy lists a different entity, “Expert Center”, as the app’s holding company, which also did not respond to requests for comment.
The opaqueness of chatbots can be particularly fraught when users share sensitive information. MystiCall’s privacy policy acknowledges conversations may touch on users’ “relationships, mental state, health, sexuality, beliefs, finances”. The policy for Your AI Psychic does not address sensitive information at all.
Kellett said this lack of transparency around how data is handled makes her particularly wary of AI apps. She said a number of platforms have contacted her for help training their LLMs, but when she pushes for more details, they tend to disappear.
Unlike many workers who fear they will someday be replaced by AI, Kellett said she does not believe her ability can be replicated by chatbots. Even if she were to hand over all of her training courses for developing psychic ability to an LLM, she does not think AI is yet at a point where it could make the leap to intuition.
“Psychic ability itself is the ability to connect to a divine source, and AI cannot do that,” she said. “AI can’t pull information from somewhere else. It can only derive information from what you give it. If you ask it to meditate, to sit down, to clear its mind, and connect to a psychic source, how would it do that? It’s just a computer.”
來源:The Guardian Technology · theguardian.com