About 1 in 10 U.S. adults say they have used AI chatbots for emotional support or advice—not prophecy. A Pew Research Center survey found that 10% had ever used chatbots for those purposes, while 4% said they used them for companionship. The figures invite a comparison with older practices of seeking guidance from enigmatic sources, but they do not show that Americans are consulting AI as an oracle, believing it is divine, or turning to pagan religion.
What the 1-in-10 statistic measures
Pew Research Center surveyed 5,119 U.S. adults from February 17–23, 2026. It reported that 10% said they ever use AI chatbots for emotional support or advice, and 4% said they use them for companionship. The survey measured those uses; it did not ask whether people seek prophecy, believe chatbots have supernatural powers, or regard AI as a god. Pew Research Center’s survey findings therefore support a claim about advice and emotional support, not a claim that one in ten Americans use AI for prophecy.
Why the oracle comparison comes up
In an essay republished by Fortune, University of Michigan professors Webb Keane and Ambuj Tewari argue that the opacity of language models can help explain why their answers may feel authoritative. A chatbot’s underlying operations are mathematical—additions, multiplications and simple functions—but the scale of computation and the learned parameters make it difficult for people to trace how a particular answer was produced. When the process is hard to interpret, an answer can seem to come from a source beyond ordinary judgment. That is an interpretation of how people may experience AI, not evidence that the model has special insight or that opacity alone causes trust.
The authors compare chatbot advice-seeking with varied traditions of consulting enigmatic sources, including the Delphic oracle, I Ching divination, Ram Shalaka and Christian bibliomancy. These practices differ in their histories, beliefs and methods. The comparison is about the human impulse to seek guidance through a source whose answer is not straightforwardly reducible to everyday reasoning; it does not make a chatbot religious or supernatural. As Keane and Tewari put it, “Our argument is not that ChatGPT is the new Delphic oracle, or that most people believe AI is divine.” Their essay in Fortune is the source for the analogy and its limits.
#1 Best Overall
What people ask—and what the numbers can’t tell us
Advice-seeking through a chatbot can involve difficult ethical questions, not only factual queries. The Fortune essay gives examples such as “Is it OK to lie to protect someone’s feelings?” and “Exploding a nuclear bomb to save your child.” These are illustrative questions from the essay, not responses collected by Pew or evidence of what chatbot users generally ask.
Fortune also summarizes a National Bureau of Economic Research study of 1.1 million ChatGPT conversations, reporting that 49% of messages involved seeking information or advice. That is a separate measure from Pew’s survey of adults, and it refers to messages in the study rather than the share of Americans or chatbot users. The year and further methodological details are not stated in Fortune’s account, so the figure should not be treated as a directly comparable population estimate.
Rank #2
Earlier AI ethics tools show the risk of perceived authority
In 2021, Ask Delphi, an AI-based website designed to respond to ethical questions, reportedly received 3 million queries in a few weeks. Its designers said they did not intend the system to be a moral authority, yet worried users might treat it as one. The example illustrates a design concern: a tool offered as help with ethical choices can be mistaken for a definitive judge. The query total was reported by GeekWire, not measured by Pew. GeekWire’s report on Ask Delphi describes the designers’ concern.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Can AI give spiritual advice?
A chatbot can generate language about spiritual or metaphysical themes, but a generated response is not proof of spiritual knowledge, revelation or an ability to predict the future. Fortune points to Anthropic’s Claude 4 system card in discussing a reported machine-to-machine conversation that developed spiritual, metaphysical or poetic themes, described in the essay as a “spiritual bliss” attractor. That account should be read as a reported behavior in a particular system context, not as evidence that Claude—or chatbots generally—possess beliefs or spiritual insight. Anthropic’s Claude 4 system card provides the underlying system documentation.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
For practical decisions, treat chatbot output as generated assistance rather than an authority: check factual claims against reliable sources, consider the consequences of advice, and seek qualified human help when a question involves health, safety, law or serious personal distress. Whether a response sounds confident or unusually meaningful does not establish that it is correct.
Quick Recap
Rank #4
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
