"The chat is my best friend." What do we do when teenagers turn to artificial intelligence to ease their loneliness?
While adults are convinced they’re using it to write school assignments, teenagers are actually sharing their deepest emotional struggles with artificial intelligence — social anxieties and moments of depression. What are the dangers of exposing personal information to a machine? How can we create an alternative outside the screen? And how can AI still be used safely?
Written by Kobi Rubinstein
Ynet
May 10, 2026
At a time when we are all focused on the importance of having a responsible adult present in teenagers’ lives as one of the ways to stop the rising violence, questions are also being raised about how teenagers use artificial intelligence. According to a recent study conducted at the University of Haifa, more than 80% of teenagers in Israel turn to AI-based chatbots for social and emotional support on issues such as loneliness, anxiety, conflicts with family and friends, or depression.
When teenagers turn to a chatbot, they are looking for a space that feels safe — a place without judgment, without fear of being reported, and without a social price for emotional conversations. "In the past, teenagers turned to Dr. Google. Today, they turn to Professor ChatGPT. They say, 'The chat told me,'" says Maya Magnat, Director of Digital Spaces at ELEM Digital. She adds that "teenagers even share their conversations with artificial intelligence on social media as a way of finding others like them."
In certain cases, a connection with a chatbot can truly be meaningful for teenagers and sometimes even help them. Corresponding with the chat can feel comforting and serve as a kind of safety net that helps them cope with loneliness, allows teenagers to feel safe talking about things they would not dare discuss face to face, or receive information on subjects they are embarrassed to bring up or simply have no one to ask.
But a relationship like this can become a red flag when the chatbot becomes too large a part of a teenager’s life — or even their only social connection. Warning signs include when a boy or girl has difficulty making decisions without the chat’s help, spends all day messaging with the chat while giving up other activities in order to talk to it, or believes it is a real being with emotions and consciousness, and that everything the chat says is true.
Another problem is that chatbots can sound very convincing, but the information they provide is not always accurate or reliable. AI systems are based on information drawn from the internet, which is sometimes outdated or untrustworthy. The chat also tends to agree with the user and reinforce what they already think or feel, instead of challenging them or offering a more nuanced perspective. For teenagers dealing with emotional distress, this can be especially problematic.
But there is still a lack of awareness about the issue among parents and those who work with teenagers. "AI tools are becoming part of teenagers’ emotional world," explains Magnat. "Adults are mainly focused on the implications of AI in the context of schoolwork. We see the impact of AI on teenagers in the field, but there still isn’t enough awareness of it. It’s important to understand that this is happening — and if it’s happening, we need to talk about it."
One organization trying to provide a different kind of response is ELEM/Youth in Distress through ELEM Digital, where they assist teenagers and young adults in distress online, operating with support from the National Insurance Foundation funds. In the outreach unit, professionals and volunteers scan TikTok and Instagram to identify teenagers in distress and reach out to them.
The organization also operates an online treatment center, in cooperation with the National Insurance Foundation funds and Hotline 105, which handles, among other things, cases related to harm carried out using artificial intelligence. Over the past two years, through the outreach unit, the online treatment center, and the chat support system, the professionals at ELEM Digital have connected with around 20,000 teenagers in the digital space.
One of the cases identified by the outreach unit was that of N. A few months ago, the outreach team noticed troubling videos that N. had uploaded to TikTok about eating disorders and anxiety, and they contacted her. After a gradual process of building trust, during which a volunteer sent her messages, expressed concern, and invited her to talk, N. began communicating with them.
Following the exchange, the team understood that N. was genuinely dealing with significant difficulties and wanted to make changes in her life. She also expressed willingness to receive support outside the internet. As a result, a treatment coordinator from one of the organization’s holistic centers contacted her personally in order to ease the initial meeting. N. agreed and began attending regular sessions at the youth center, where she met other teenagers and formed a meaningful connection with the coordinator, who helped her cope with her difficulties.
Unlike AI, when teenagers turn to ELEM Digital, they encounter real human beings on the other side of the screen. Growth emerges from the human connection between the volunteer and the teenager. This human relationship — with its misunderstandings, disagreements, and frustrations — can help a boy or girl learn to see their story from a different perspective, put themselves in someone else’s shoes, or recognize that there are other possible courses of action they had not considered.
“They ask if we’re bots.”
Another tool offered by ELEM Digital is an anonymous support chat for teenagers, in which volunteers with therapeutic backgrounds provide emotional support to teens and young adults. Support is available in Arabic on Sundays, Tuesdays, and Thursdays from 8 p.m. to midnight, and in Hebrew on Sundays through Wednesdays from 8 p.m. to midnight.
Magnat says that over the past year, one question has opened almost every conversation: are you a bot or a human being? That question reflects just how accustomed teenagers have become to discussing personal and emotional issues with AI systems.
"Because they’re used to talking with chatbots, they are more impatient — sometimes even impolite — and they relate to the volunteers in ELEM Digital’s chat somewhat like they would to a machine. It’s important for professionals to recognize this change and think about how to deal with it and respond to it when working with teenagers," Magnat explains.
At the same time, there is also an encouraging sign. "I think there is a shift in the policies of the companies developing these chat systems," says Magnat. "They understand that there are problems and responsibilities on their side, and one of the things they are trying to do is have the chat recommend turning to a human and professional source of support at a very early stage."
So what can still be done? "It is very important to have conversations with teenagers about the issue — not from a judgmental place, but from a place of open, accepting, and genuinely interested dialogue," says Magnat. "To talk about what AI is, how it works, whether what the chat says can be trusted, and which topics are appropriate to discuss with it and which are better discussed with a human being."
At a time when cases of violence among teenagers are increasing, it is even more important to create opportunities for real human conversation. When teenagers become accustomed to a space with almost no resistance, disagreement, or friction, encounters with real people can start to feel more difficult, frustrating, or unsatisfying. Teenagers need spaces where they can talk, hesitate, and even get angry — while also receiving boundaries, guidance, and tools for coping with conflicts and meeting their needs. In a period when social reality is becoming more tense and violent, the importance of human support and the presence of meaningful adults in teenagers’ lives is only continuing to grow.
What should we know and communicate to teenagers?
Magnat emphasizes several important points. First, AI does not truly know us. "The goal of the chatbot is to please us," Magnat explains. "It’s a statistical machine that doesn’t really know who we are and doesn’t understand the context of the questions."
It is important to remember that the information provided by AI models is based on freely available information from the internet, which is not always reliable.
In addition, when teenagers rely on the chat to analyze situations for them and decide how to act, there is concern that their ability to make decisions independently may be harmed. Just as we have become accustomed to relying on navigation apps and forgotten how to navigate on our own, the ability to make independent decisions could also weaken.
"As long as the teenager understands the limitations, knows how it works, and occasionally talks with it to gain another perspective, it can be used," Magnat emphasizes. "But when teenagers feel that the chat truly knows them, begin to depend on it, and can no longer make choices on their own — that is very concerning."
And most importantly — basic safety rules must be maintained: not using artificial intelligence to harm others, such as by creating or sharing embarrassing images. "And of course, to discuss the most important question of all — who in our lives can we turn to when something frightening, disturbing, or strange happens," says Magnat.
Looking at ourselves — the parents and adults who interact with teenagers — Magnat concludes by saying, "We have a tendency to think that technology can solve all problems. But we cannot rely on the chat to take over our role in providing emotional response. If anything, what we need to ensure is that our response to teenagers becomes more accessible."