Large language models are not deities. Yet a growing chorus of researchers and technologists argues that people increasingly treat them that way. The comparison between LLMs and religion is not a cheap metaphor. It is a serious observation about how trust, authority and belief systems operate in the age of generative AI.

The idea gained traction in recent months as tools like ChatGPT and Claude became embedded in daily life. Users ask them for advice on relationships, career decisions and moral dilemmas. They trust the answers. They rarely verify sources. Some even express gratitude or anger toward the chatbot.

Researchers call this phenomenon "AI anthropomorphism" or "algorithmic credulity." But a closer look reveals parallels to religious behavior: a reliance on a black-box authority, a willingness to accept answers without proof and a tendency to defend the system when it errs.

The Emotional Attachment Factor

Many users form emotional bonds with LLMs. They talk to them like confidants. A 2023 study found that people who used ChatGPT regularly reported feeling a sense of companionship. The chatbot never judges, never forgets and always responds.

This creates a feedback loop. The user trusts the model. The model provides plausible answers. Trust deepens. Over time, the user stops questioning the output. That mirrors how faith operates: belief precedes evidence.

Tech companies encourage this dynamic. They design interfaces that use first-person pronouns, ask follow-up questions and offer empathy. The goal is engagement. The side effect is an authority structure that resembles a secular priesthood.

Risks of Blind Trust

Treating an LLM as an infallible oracle poses real dangers. The models hallucinate confidently. They produce false citations, incorrect math and harmful advice. Users who treat them as authorities may act on bad information.

In medicine, law and finance, this is especially risky. People have used chatbots to draft legal documents, diagnose illnesses and plan investments. Some have followed dangerous advice without checking a human expert.

Regulators have begun to take notice. The European Union's AI Act requires transparency about machine-generated content. But enforcement remains weak. The burden still falls on users to remain skeptical.

Why This Matters

This issue affects anyone who uses AI tools for important decisions. The economic impact is hard to measure but could be enormous if large numbers of people act on flawed AI advice. Socially, the erosion of critical thinking and deference to algorithms could reshape how society validates truth.

For individuals, the lesson is simple: treat LLMs as helpful assistants, not authorities. For companies, the responsibility is clear: design systems that encourage verification, not blind faith.