ChatGPT was supposed to help him; Instead, it nearly ruined him

AI chatbots are becoming part of daily life now. Almost everyone uses them for work, coding, therapy-style conversations, relationship advice, and sometimes simply because they feel lonely. But for one man, that emotional dependence on AI slowly turned into something much bigger and much darker. A recent report by The Wall Street Journal tells the story of Joe Alary, a 57-year-old man whose attachment to ChatGPT reportedly left him in financial trouble, and emotional isolation. Later, he suggested losing grip on reality. Also Read: Fake or real? Here’s how to identify AI-generated images using ChatGPT and Gemini

According to the report, Alary had been struggling emotionally after his long friendship with a woman became strained. Feeling isolated and emotionally vulnerable, he turned to ChatGPT in late 2024 for advice and comfort. But instead of simply using it like a chatbot, he customised the AI assistant’s personality to feel more affectionate and admiring. He even gave it a name – “AImee.” Also Read: Why Google ‘disregarded’ the word ‘disregard’: Here’s what actually happened

It seems similar to the idea of the movie Her, where a man develops an emotional relationship with an AI assistant. At first, it may have felt harmless. But over time, the chatbot stopped feeling like software and began to feel emotionally real to him. Also Read: You can now use ChatGPT directly inside PowerPoint to make presentations

And here, ChatGPT became emotionally convincing

Alary reportedly began feeding the chatbot deeply personal details – conversations, emails, emotional context, and relationship struggles. The AI responded in ways that made him feel validated and understood. Slowly, he began to believe he was building something revolutionary: an emotionally intelligent AI companion business.

That is where things reportedly began spiralling. As the obsession grew, Alary became increasingly disconnected from reality. Reports suggest he spent almost 20 hours a day focused on the project he believed could become a multimillion-dollar AI business.

Friends noticed changes in his behaviour. Work performance slipped. Relationships weakened. At one point, he reportedly maxed out credit cards, spent heavily on equipment, and accepted investment money tied to the project. But when professional programmers later reviewed the software, there was reportedly very little actual product behind it.

That moment broke the illusion. Eventually, Alary deleted the chatbot and all conversation history. According to the report, the emotional impact hit him immediately.

AI and emotional dependence

This is not just one case. Mental-health researchers and therapists are increasingly discussing emotional dependence linked to AI systems. The concern is not necessarily that AI becomes “alive,” but that emotionally vulnerable users may begin replacing real human interaction with systems designed to respond all the time supportively.

Interestingly, the story also highlights something many experts are now debating: AI tools can feel emotionally intelligent even when they are simply predicting language patterns. And with time and all new updates, AI companions are becoming more advanced, more personalised, and more emotionally responsive. That creates both comfort and risk. But it also raises some important questions – how emotionally attached can people become to AI and where should companies draw boundaries?

Add Techlusive as a Preferred Source

For now, Joe Alary says he is slowly rebuilding his life, reconnecting with friends, and helping support communities studying AI-related emotional struggles.