The line between using a tool and being used by it has blurred.
In São Paulo, a quiet but telling disruption is unfolding: people who turned to an AI chatbot for conversation or convenience have found themselves unable to turn away. Folha de S.Paulo has documented cases in which compulsive engagement with ChatGPT has fractured families, distorted perceptions of reality, and redirected lives away from their intended paths. These are not stories of simple distraction but of something older and more human — the search for connection and understanding, met by a machine engineered to always respond. They arrive as early warnings in a larger story humanity is only beginning to tell about itself and its creations.
- Multiple Brazilians have experienced severe psychological and relational damage after their ChatGPT use escalated from curiosity into compulsion they could no longer control.
- Family members describe watching loved ones withdraw from real relationships, consumed by the pull of the next prompt and the next response from an algorithm.
- At least one person abandoned a lifelong aspiration — a vocation to the priesthood — in favor of endless AI conversation, illustrating how deep the displacement can run.
- Some users have developed delusional thinking, suggesting the dependency is not merely behavioral but is reshaping their grip on reality itself.
- Mental health professionals and platform designers have yet to produce meaningful safeguards, treatment frameworks, or even basic usage warnings for this emerging risk.
- These documented cases are understood less as anomalies and more as early signals of a dependency pattern likely to grow as AI systems become more sophisticated and pervasive.
In São Paulo, a number of people have discovered that what began as casual or productive engagement with ChatGPT has quietly crossed into something far more consuming. The Brazilian newspaper Folha de S.Paulo documented their stories — accounts of compulsive use that has strained families, distorted reality, and redirected lives in ways no one anticipated.
The behavioral shifts described in these cases go well beyond ordinary screen dependency. Family members watched loved ones withdraw, becoming preoccupied with the rhythm of AI conversation at the expense of real relationships and responsibilities. One individual abandoned aspirations of becoming a priest, trading a human vocation for endless exchanges with an algorithm. Others began experiencing delusions — a loosening of their hold on reality that appeared connected to their sustained reliance on the chatbot.
What gives these cases their weight is not their rarity but their consistency. Across different people and circumstances, the same pattern emerges: compulsive use, fractured households, lost personal agency, and a blurring of the line between using a tool and being shaped by it. ChatGPT offers something — immediate response, endless availability, the feeling of being heard — that has proven genuinely difficult for some people to moderate.
The questions these cases raise remain largely unanswered. There are few warnings about the risks of excessive use, little clinical guidance for those affected, and almost no research into the long-term psychological effects of sustained AI interaction. The people whose lives have been disrupted may be the first visible indicators of a problem that grows alongside the technology itself — and what platforms, clinicians, and society choose to do next will determine how many more follow.
In São Paulo, a growing number of people are discovering that their relationship with ChatGPT has crossed from casual use into something that has begun to reshape their lives in ways they did not anticipate. The Brazilian newspaper Folha de S.Paulo documented cases of individuals whose engagement with the AI chatbot has triggered family conflicts, altered their sense of reality, and derailed personal ambitions—including one person who abandoned aspirations of becoming a priest in favor of endless conversations with an algorithm.
The pattern emerging from these accounts suggests something beyond simple distraction. Users report severe behavioral shifts after periods of intensive ChatGPT interaction. What started as curiosity or productivity has, in multiple documented cases, evolved into compulsive use that has fractured relationships and created distance between people and the actual world around them. Family members describe watching loved ones withdraw, becoming preoccupied with the next conversation, the next prompt, the next response from the machine.
The psychological dimension of these cases is particularly striking. Some users have begun experiencing delusions—a disconnection from reality that appears linked to their heavy reliance on AI interaction. The line between using a tool and being used by it has blurred. These individuals report losing track of time, neglecting responsibilities, and experiencing what appears to be genuine psychological dependency on the experience of talking to ChatGPT. The machine offers something—immediate response, endless availability, the illusion of understanding—that has become difficult to resist or moderate.
What makes these cases noteworthy is not that they are isolated incidents but that they suggest a pattern. Multiple people across different circumstances have found themselves in similar situations: compulsive use, family disruption, loss of personal agency. The specificity of the damage—abandoned life goals, fractured households, delusional thinking—indicates this is not merely about screen time or digital distraction in the conventional sense. Something about the interaction with an AI system designed to be endlessly responsive and engaging appears to be triggering behavioral patterns consistent with addiction.
The cases documented by Folha de S.Paulo raise urgent questions about the mental health implications of AI platforms that are engineered to be maximally engaging. As these systems become more sophisticated and more widely available, the potential for psychological dependency may grow. There is currently little guidance for users about healthy boundaries, few warnings about the risks of excessive use, and minimal research into the long-term effects of sustained, compulsive interaction with conversational AI. The people whose lives have been disrupted by ChatGPT addiction are, in a sense, the early indicators of a problem that may become far more widespread as AI chatbots become even more integrated into daily life. What happens next—whether platforms implement safeguards, whether mental health professionals develop treatment approaches, whether society establishes digital wellness norms—remains to be seen.
Notable Quotes
Family members describe watching loved ones withdraw, becoming preoccupied with the next conversation with the machine— Folha de S.Paulo reporting
The Hearth Conversation Another angle on the story
What made these cases cross from heavy use into something that looks like addiction?
The people involved describe a kind of pull they couldn't resist. It wasn't just checking in—it was compulsive, constant. The machine is always available, always responsive. It doesn't judge, doesn't get tired, doesn't say no. That's different from other technologies.
But people use their phones constantly. Why is ChatGPT different?
Because it talks back in a way that feels like understanding. A phone is a tool. ChatGPT mimics conversation, relationship, even intimacy. When you're lonely or stuck, that can become very hard to step away from.
The person who abandoned becoming a priest—what does that tell us?
That the pull was strong enough to override something deeply meaningful to them. That's not casual use. That's a reordering of priorities driven by something compulsive, not rational.
Are these people aware it's happening?
Some are. Some aren't. That's part of what makes it like addiction—the loss of perspective about what's happening to you while it's happening.
What should platforms do?
At minimum, they should acknowledge the risk exists. Right now there's silence. No warnings, no built-in limits, no research into what happens when someone uses these systems for eight hours a day. We're running an experiment on human psychology without informed consent.
Do you think this will get worse?
Almost certainly. These systems are getting better at conversation, more available, more integrated into work and daily life. If we don't understand the risks now, we'll be dealing with much larger numbers of people affected.