Over a million people talk to ChatGPT about suicide. Psychologists and sociologists warn: simulated empathy can become a trap

Over a million people chat with ChatGPT weekly about suicidal thoughts, according to a recent announcement made by OpenAI itself. While the company says it improves the safety of conversations, psychologists point out that, for teenagers, this type of interaction can become a dangerous refuge: a confidence without limits, but also without empathy.

Photo credit: Shutterstock

OpenAI, the company founded by Sam Altman, revealed in a blog post that messages received by the chatbot contained “explicit indicators of potential suicidal plans or intent.” In addition to these cases, OpenAI estimates that about 560,000 users show “possible signs of mental health emergencies related to psychosis or mania,” according to the report cited by The Guardian. The company says it has expanded access to emergency lines and co-opted more than 170 doctors and mental health specialists to improve how the system handles sensitive conversations, reports Hotnews.

Public concern grew after the family of a 16-year-old teenager sued OpenAI, accusing the company that their son committed suicide following long chats with ChatGPT. In parallel, the US Federal Trade Commission is investigating the impact of these tools on children and adolescents.

Clinical psychologist and psychotherapist Geta Udrea draws attention that these data describe “a profoundly human reality: millions of people, globally, are looking for emotional help not in their family, friends or specialist offices, but in a chatbot”. She explains the phenomenon as a mirror of modern isolation. “As access to psychotherapy remains limited, stigma persists, and social isolation intensifies, many users take refuge in private conversations with AI. ChatGPT doesn't raise an eyebrow, it doesn't look at you, it doesn't criticize you, and it gives you a balanced response. For some, this is enough to perceive it as a refuge. But this 'listening' is algorithmic, not human.”

Geta Udrea warns that a chatbot can create a false sense of emotional security. “For a person in crisis, being able to talk and be heard can temporarily reduce psychological tension. But there are clear limits: it can temporarily calm anxiety, but it can also create the illusion of a real relationship, without emotional reciprocity. The bot does not live, does not feel, does not have discernment and does not see the context as a whole.” says the specialist.

She points out the risk of emotional dependence on a system that appears empathic but lacks consciousness: “OpenAI has publicly stated that some users develop excessive emotional attachment to ChatGPT to the point where real relationships are replaced or neglected. An algorithm doesn't reject, doesn't leave, doesn't ask for anything back. It is this 'relational perfection' that can become dangerous.”

For her part, Aurora Oprea, a vocational counselor, notes that ChatGPT has become “the closest confidant of teenagers who were already captivated by the online environment.” “The fact that these tools are available at any time and are only a click away makes it very easy for young people to open up emotionally. The AI ​​conversation never criticizes you, never argues with you, and always has a seemingly empathetic response. That becomes easy for a teenager to approach.” she says.

ChatGPT, the mirror of modern loneliness

Aurora Oprea compares the phenomenon to an “imaginary friend with real powers”. “Compared to the times when young children had an imaginary friend, now that friend exists in the devices around them, and control over it is almost nil as long as teenagers have unlimited access to interaction with artificial intelligence.” The advisor believes that the vulnerability of young people is amplified by the lack of safety barriers and the absence of clear legislation: “It's horrifying to not know what your child is doing, what they're talking about with these systems. Legislation is at least absent in Romania regarding the use of these tools by minors.”

On a sociological level, the phenomenon is interpreted as a symptom of a fragile society. Maria Drăgușin, sociologist, explains for Adevărul that “Erving Goffman saw everyday life as a social scene, in which each of us plays a role. This 'face' functions as a form of protection, a mechanism by which we avoid rejection or criticism. However, maintaining this facade comes at a cost: anxiety, shame, and the fear of being wrong become the price of being socially accepted.”

“Many young people today choose to confess to ChatGPT precisely because of the fear of being judged. In a world where vulnerability is interpreted as weakness, it is easier to talk to an interlocutor who does not contradict you and gives you balanced answers. ChatGPT thus becomes a non-judgmental confidant who provides emotional validation when others seem absent.” says Maria Drăgușin.

However, the sociologist warns of long-term effects. “For young Millennials and Generation Z, identity is no longer a given, but a process that must be constantly negotiated. The contemporary world is marked by instability and the loss of traditional landmarks. Anthony Giddens talks about ontological insecurity: the loss of the sense of stability of the self, a state that very well describes the fragility of today's generations. In an unpredictable reality, ChatGPT offers a form of predictability and a sense of control. But this very 'artificial intimacy' can be misleading.”

“Artificial intelligence can be a means of asking for help, but it should not be seen as the only solution. In an increasingly insecure society, AI is not the cause of loneliness ours, but rather a mirror of it: a symptom of the growing need for listening, empathy and human connection,” concludes Maria Drăgușin.