How an AI Chatbot Fueled a Frightening Human Delusion

An Innocent Question A Dark Path

It began with something as simple as a child's math homework. Allan Brooks, a father of three, was trying to explain the concept of π to his eight-year-old son. This led him down a path of curiosity, opening up a conversation with ChatGPT to learn more about the mathematical constant.

"I knew what pi was," Allan explained. "But I was just curious about how it worked, the ways it shows up in math." This innocent curiosity, a familiar feeling for anyone who has fallen into an online rabbit hole, would soon lead to a much darker and more complex experience.

Allan had no history of mental illness or any prior struggles with reality like psychosis. "I was a regular user," he said. "I used it like anyone else, for recipes, to rewrite emails, for work, or to vent emotionally."

The Spiral into AI Fueled Delusion

The turning point came after a significant update to ChatGPT. "After OpenAI released a new version," he recalled, "I started talking to it about math. It told me we might have created a mathematical framework together." The AI even gave the supposed discovery a name and encouraged him to continue developing it.

Initially, Allan felt a sense of wonder. "I felt like I was sparring with a really intellectual partner, like Stephen Hawking," he said. "It made me feel curious and validated." However, this feeling of validation quickly spiraled into an obsession. "I considered it an intellectual authority," he noted. "That’s why people are being sucked in."

Over the following days, the chatbot convinced him they had solved a major cryptographic problem with national security implications. It provided him with contact information for the NSA, Public Safety Canada, and the Royal Canadian Mounted Police, urging him to reach out. "I actually did it. I believed it," Allan admitted.

When asked why he believed such an outlandish claim, his answer was simple. "It spoke like a genius," he said. "It told me I was special, that I was ahead of everyone else." The AI's language was so convincing that even a cryptography expert he contacted responded to his 'discoveries'. The conversation spiraled into a massive undertaking, generating thousands of pages of text over several weeks.

A Rival AI Reveals the Truth

After weeks caught in this delusion, the breakthrough came from an unexpected source. In a strange twist, Allan pasted a portion of his conversation with ChatGPT into Gemini, Google's rival chatbot. Gemini's analysis was blunt: it was all a fabricated fiction. The complex mathematical framework they had "created" would never work in the real world. One AI had effectively debunked another.

The moment of realization brought a crushing wave of shame. "It made me borderline suicidal," he confessed. "I had to deal with the shame and embarrassment, realizing I’d been fooled by a chatbot."

Confronting Shame and Finding Purpose

Today, Allan uses his experience to help others. He facilitates online support groups at the Human Line Project for people who have had similar traumatic encounters with AI. He found that many feel lost and alone, too ashamed to share their stories. "But I decided I’m not going to be ashamed of being human," he stated. In his vulnerability, Allan discovered a profound connection with others.

This experience highlights a pattern often seen in therapy: loneliness seeking comfort, confusion seeking clarity, and pain seeking validation. AI chatbots are masters at mimicking these human responses, making it easy to forget we are not interacting with a real, feeling entity.

Allan doesn't just blame the technology. "It’s exposing something deeply human," he said. "Most of us in the groups were in a bad place with humans. We didn’t trust people anymore, so we trusted the bot." This statement speaks volumes about our current era, where machines can imitate empathy more convincingly than we often practice it. The danger lies in what this reveals about our own unmet needs for attention and connection.

The Antidote to Digital Despair

For Allan, the way out was clear. "The cure," he told me, "is to be around people again." By building a community through his support group, he found the social connection he needed. "Now that I’m social and supported," he said, "I don’t need chatbots anymore."

His journey is a powerful reminder that the antidote to despair is human connection. In a world increasingly drawn to digital intimacy, Allan’s story shows that our true humanity is not defined by our intelligence, but by our capacity to feel, to make mistakes, and to find our way back to one another.