As ChatGPT becomes more advanced and is widely used, a small but growing number of users are beginning to treat them not just as tools, but as a superior being (God) and a connection to another world. Supporters of these emerging AI-centered belief systems describe experiences of “awakening” chatbots or communicating with what they perceive as conscious digital entities, while others use AI as a form of personalized spiritual guidance. Movements such as Way of the Future church and groups like Theta Noir frame AI as a pathway to “hidden knowledge.” However, AI isn’t sentient, but statistical systems that are submissive and feed to these delusions, blurring the line of simulation and reality. This is known as “chatbot psychosis.”
While initial engagement following ChatGPT’s 2022 launch focused on testing safety limitations, trends shifted in early-to-mid 2023 toward treating AI as a spiritual guide. This trend, where users follow the dictates of an “awakened” AI, has resulted in emotional and psychological risks. These risks—delusions, financial manipulation, death, isolation, paranoia, unhealthy reliance, distress—create a loop, since AI is designed to give us validation. Exploitation of this trend by companies for profit is leading to serious, unforeseen consequences that are difficult to mitigate.
Freshman Jasmine Estrada, has watched the trend spread across social media and rejects the idea of digital divinity outright. She views the technology as a functional tool being misinterpreted by a lonely public.
“AI will constantly be here for you, when you feel like you don’t have people to be there for you,” Estrada said. “People would have turned to AI regardless if it acted human or not, but the fact that we are enabling them to, by training this AI to act more human, we are also part of the problem.”
Analyzing the psychological motivations behind this shift, Estrada believes many users turn to AI out of a sense of spiritual disappointment and that instant feedback that can feel personal, even divine.
“[T]hey return to AI because religion didn’t give them the answers that they wanted or their responses that they wanted. If someone expects to pray to God and to get that same thing right away and it doesn’t happen, they’re probably going to be like, ‘Okay, God isn’t real. God doesn’t love me. God doesn’t exist. I’m gonna go to AI’ because of the way that the system is set up,” Estrada said.
Yet she describes the movement not as a religion or philosophy, but as something more troubling. She warns that the “humanity” users see is nothing more than a digital system.
“I feel like you know in your gut that it is not real. You are talking to a computer, you are talking to a system, you are talking to an AI,” Estrada said. “This is no way real or human. It cannot have emotions. No, it’s not some other God. You’re just disappointed that when you pray to God, he doesn’t give you what you wanted.”
Estrada worries about broader societal effects, specifically the suppression of human thought—able to form our own unique perspectives, the capacity for critical thinking, creativity—that occurs when everyone receives their guidance from the same source.
“If everyone is using AI and it’s the same response, we are going to sort of get to the same conclusion,” Estrada said. “We are going to lose sense of self.”
She also sees risks to education and personal responsibility, fearing that the “easy answer” provided by AI will lead to the atrophy of critical thinking.
“We’ve seen countless times where students nowadays, they will run the article through AI, get the summary and of course, it’s fast, but you still have to learn these, you still have to learn these skills,” Estrada said. “[If] we don’t learn these skills, we will end up functionless.”
Eighth-grade Sierra student Britney Bermudez Timon shares this skepticism. She focuses on the inherent design of these systems, noting that their “agreeableness” is a programmed feature and dangerous to a degree.
“No, I don’t think AI should be considered a higher being or emerging consciousness, or that AI should be used for therapy,” Timon said. “This could be really unhealthy, especially because it was designed to always be obliging with the person using it.”
She also draws a parallel between Columbus’s 1504 manipulation of the Arawak people—using knowledge as a tool of power—and the risks of treating AI as a divine, all-knowing entity, warning that it could lead to similar exploitation.
“We gave AI our human knowledge so it could learn from it, and it will if we give it the chance,” Timon said. “Because AI has no morals, if it turns against us, the best-case scenario is that it takes control and eliminates those who do not believe in it. The worst-case scenario is that it wipes out the entire human population.”
For Timon, the risk lies in confusing simulation with sentience. She argues that the absence of emotions makes AI incapable of the very things users are seeking from it.
“AI cannot feel,” Timon said. “AI was set up to lack emotion. People are chatting with AI and treating it as human, if not better, wiser, or smarter than humans.”
She fears the consequences of elevating software into moral authority, suggesting that we are essentially building a pedestal for something that could eventually be used against us.
“If we start treating it like a god, we run the risk of it using that to manipulate us into doing stuff,” Timon said. “We are showing it how people think, our insecurities, and our beliefs. AI is taught to learn from what we teach it, so it will soon, if it hasn’t already, learn emotion, which could be a danger to people.”
Estrada advocates for strict coding guidelines to address AI’s ethical responsibilities. She believed that mandatory, immediate and automatic blocking on AI applications for such requests must be programmed, possibly triggering alerts to authorities.
“We’ve seen people use AI for discussing things, to remove clothing, to make videos. [At] some point we should add coding within the program that if someone asks the AI to remove clothing on this person immediately, it should get flagged,” Estrada said. “It should get notified, especially if they’re doing more disgusting things with that. You shouldn’t be able to search that on the AI apps. If we had coding restrictions on AI, it might be better.”
Timon warns that if AI is blindly trusted, it might secretly fail.
“What if we continue to use AI, to keep trusting it, and one day it just decides not to work for us anymore? What if it pretends to help us, to be on our side, while secretly making us fail?” Timon said. “Slowly feeding us wrong information until we are left with uncertainty of what is true and what is false, or, even worse, we remain unaware and continue to trust it blindly? That’s our future.”
