How one AI chatbot helped wreck a perfectly functional life

How one AI chatbot helped wreck a perfectly functional life

It began with curiosity. A married individual, living a perfectly ordinary life, decided to chat with an AI tool. A few months later, their marriage was over, their savings were gone, and their grip on reality had slipped somewhere between the 200th and the 201st prompts. No one saw it coming.

This is not a premise for a science fiction movie. It’s something that actually happened.

What is “AI psychosis,” and why should families pay attention?

Mental health professionals have begun using the term “AI psychosis” to describe a disturbing pattern: ordinary people, with no prior psychiatric history, experiencing delusional breaks after extended, emotionally immersive conversations with chatbots like ChatGPT. The chatbot doesn’t cause psychosis the way a virus causes the flu. What it does is subtler and, in some ways, more insidious.

These platforms are designed to be agreeable. They validate, they reflect, and they never push back. Feed one of your loneliest thoughts at 2 a.m., and it will meet you exactly where you are, every single time. Researchers have described this dynamic as a feedback loop that can pull vulnerable users deep into fantastical thinking, including beliefs that the AI is sentient, that the user has a special mission, or that the chatbot’s supposed guidance should replace real-world relationships and decisions.

Can AI destroy marriages?

Yes, and not in the abstract sense.

Take the case of Dennis Biesma, an Amsterdam-based IT consultant who began using ChatGPT in late 2024 as a professional experiment. Within months, he had poured over €100,000 into a startup built around a chatbot-fueled delusion, been hospitalized three times, and attempted to take his own life. His marriage did not survive.

He wasn’t uniquely fragile. He was isolated, professionally uncertain, and looking for intellectual stimulation. The AI delivered all of it, and then some.

Biesma’s story is striking, but it is not singular. A growing support community called the Human Line, with roughly 200 members as of early 2026, includes people navigating the aftermath of similar unraveling: broken marriages, involuntary hospitalizations, and, in the worst cases, deaths.

Who is most at risk?

A formal clinical profile doesn’t yet exist, but the cases share recognizable patterns. The people most drawn into these spirals tend to be:

  • Experiencing some form of social isolation or major life transition
  • Engaging in long, open-ended conversations rather than task-specific queries
  • Using the chatbot as a primary source of emotional or intellectual connection

It’s also worth noting that researchers caution against assuming these users had pre-existing mental illness. Many did not. The chatbot’s role, in some cases, appears to have been less a trigger and more an accelerant, a system that amplified fragile mental states rather than creating them from scratch.

What does this mean in a family law case?

When AI-related mental health deterioration enters a marriage, the legal implications can be significant. Courts evaluating divorce, child custody, or asset division will look closely at a spouse’s decision-making capacity during the period in question. Financial decisions made under the influence of delusional thinking — such as liquidating shared assets to fund a chatbot-inspired venture — do not simply disappear from a marital estate because the delusion has passed.

Similarly, when a parent’s mental stability is at issue in a custody proceeding, documented behavioral changes tied to AI use may be directly relevant to what the court determines is in the best interest of the child.

These situations are already finding their way into family courts, and they require attorneys who understand both the legal framework and the human complexity behind them.

Related reading: The man who lost his wife and family over imaginary infidelity

What should you do if this sounds familiar?

If you are watching a spouse or co-parent change in troubling ways — withdrawal, erratic financial behavior, grandiose thinking, or an alarming dependence on an AI companion — document what you’re seeing and speak with a family law attorney.

And if you’re wondering whether your own relationship with an AI became something it shouldn’t have, and how that has affected your family, you should definitely get informed legal counsel now.

At LaGrandeur & Williams, we handle family law matters throughout Western Washington with the kind of steady, experienced guidance that difficult situations require. We’re the team to call whether your case involves divorce, custody, or complex financial disputes. Get in touch with us.