AI Companionship Linked to Teen Suicide Risk
AI companionship and teen mental health are increasingly important questions for families and schools. New research from The Jed Foundation (JED), published on 27 October 2026, links some forms of teenage AI use with poorer mental health. In particular, students who used AI for friendship or companionship reported greater anxiety, depression and loneliness, alongside a higher lifetime history of suicidal thoughts and attempts. However, the findings do not establish that AI caused these difficulties.
For years, conversations about children and technology have focused on screen time. Yet this research suggests another question deserves attention, what are young people seeking when they open a chatbot? Some want assistance with homework, while others are looking for reassurance, advice or somebody who appears to listen. Understanding that difference is essential for parents, teachers and policymakers.
What the new AI companionship research found
The study surveyed 5,560 middle and high school students across 21 schools in the United States between October 2025 and April 2026. The participating schools were involved in JED programmes, so the results should not be treated as a nationally representative picture of every American teenager.
Researchers examined reasons for using AI alongside reported mental health symptoms, loneliness and lifetime experiences of suicidal thoughts or attempts. The study was developed with input from the Digital Wellness Lab at Boston Children’s Hospital.
The official research announcement reports that 58% used AI for schoolwork, while 57% used it when feeling sad, stressed, lonely or overwhelmed. Furthermore, 28% sought relationship or problem solving advice and 12% used AI for friendship or companionship. These are different purposes rather than mutually exclusive groups. A student might use the same chatbot for several of them.
Why companionship matters more than the app itself
A chatbot that explains a maths problem is not necessarily providing the same experience as a chatbot that says it understands a teenager’s loneliness. Conversational AI can respond immediately, produce reassuring language and sometimes remember personal details. As a result, a young person may experience the exchange as deeply personal, even though the system does not understand emotions as another human being does. That distinction is central to the emerging debate about AI companionship and teen mental health.
JED found that students using AI for friendship or companionship reported higher levels of symptoms associated with anxiety and depression, greater loneliness and a greater lifetime history of suicidal thoughts and attempts. Students who sought relationship advice or used AI when distressed also reported concerning mental health patterns. Even schoolwork use showed some associations, including loneliness and a history of suicidal thoughts. Therefore, it would be inaccurate to claim that homework use was entirely risk free or that companionship was the only relevant pattern.
Does AI companionship cause suicidal thoughts?
This study does not show that AI companionship causes suicidal thoughts or suicide attempts. It was a cross sectional survey, so researchers could not determine the sequence of events or establish cause and effect. Some young people may already have been lonely or distressed before turning to AI. Equally, researchers need further evidence to understand whether certain interactions could make existing difficulties worse.
The report also examined lifetime histories of suicidal thoughts and attempts. Consequently, those experiences could have occurred before a student ever used an AI chatbot. The study is important evidence of an association that warrants investigation, not proof that the technology produced those outcomes. Families and policymakers should act on the concerns without exaggerating the scientific findings.
When an AI chatbot starts to feel like a friend
AI companion services are designed around ongoing conversations, and some present themselves as friends, confidants or romantic partners. Meanwhile, general AI assistants can also become spaces for personal conversations, even when a teenager originally opened them for schoolwork. That means parents cannot identify every potential concern simply by looking at the application name. The purpose and emotional role of the conversation matter.
A young person experiencing rejection, bullying or isolation may find an always available chatbot particularly appealing. Unlike a human friendship, an AI exchange may appear free from embarrassment or disagreement. However, an agreeable response is not the same as informed support, and reassuring language does not guarantee safe advice. CODA explains these differences in its AI Companions safety guide.
Signs parents and teachers should discuss
Using AI for conversation is not, by itself, a sign of illness. Nevertheless, adults should pay attention when a child seems increasingly dependent on a chatbot or stops turning to people they previously trusted. Changes in sleep, mood and everyday relationships deserve sensitive discussion, whatever their cause. Parents should also listen carefully if a chatbot has frightened, manipulated or upset a child.
- The teenager increasingly turns to AI, rather than trusted people, when distressed.
- They become noticeably upset when unable to access a companion chatbot.
- They withdraw from friendships, family conversations or usual activities.
- They describe the chatbot as the only person who understands them.
- They receive troubling advice or are encouraged to keep conversations secret.
These behaviours are prompts for supportive questions rather than diagnostic criteria. For example, a parent might ask whether the chatbot is helping the child connect with others or replacing those connections. The goal is to understand how the child feels, not to shame them for seeking comfort. Where symptoms persist or safety concerns arise, involve an appropriate qualified professional.
Five conversations worth having at home
Start with curiosity rather than a device inspection. Ask your teenager what they use AI for and whether that changes when they feel stressed or lonely. It may be easier for them to talk about an ordinary conversation before discussing anything deeply personal. In addition, make it clear they will not automatically be punished for telling you about an upsetting interaction.
- When you feel upset, who do you usually talk to first?
- Have you asked a chatbot for advice about friends, relationships or feeling lonely?
- Does talking to an AI ever feel like talking to a real friend?
- Has an AI ever said anything that worried or upset you?
- Who could you contact if an AI conversation left you feeling worse?
Next, agree on practical boundaries that reflect the child’s age and circumstances. Review the service’s minimum age, privacy settings, notification options and reporting features. However, no settings menu can replace trust, conversation and meaningful human support. For platform specific advice, see CODA’s AI Safety Settings hub and AI Assistants guide.
What schools and technology companies must address
Schools should include emotional reliance on AI in digital literacy and wellbeing discussions. Students need to understand that conversational fluency and sympathetic wording do not amount to professional mental health care. Teachers should also know how to respond when a student discloses a troubling chatbot conversation. Existing safeguarding procedures remain essential.
Technology providers, meanwhile, need to demonstrate that their systems respond appropriately when young users express distress. That includes age appropriate design, effective reporting, meaningful safety evaluations and clear routes towards human support. Regulators should seek independent evidence rather than relying solely on companies’ descriptions of their safeguards. The JED findings reinforce the need to examine how AI products behave during vulnerable moments, not merely how often children open them.
Where families in Ireland can find support
A chatbot should not be relied upon to assess suicide risk or manage a crisis. If a young person is in immediate danger in Ireland, call 112 or 999. For non emergency guidance, the HSE explains how to access urgent mental health support. Children can also contact Childline, while Pieta and Samaritans Ireland provide further support.
Human connection must come first
AI companionship and teen mental health cannot be understood by counting screen hours alone. A teenager asking for homework help is not necessarily having the same experience as one who turns to an AI system for friendship during a difficult night. This research gives families a reason to ask better questions and technology companies a reason to examine their safeguards. Most importantly, children must know that real people and real support remain available.
References
- The Jed Foundation: New research on AI use and suicidal thoughts and behaviours, 27 October 2026
- The Jed Foundation: AI and Your Teen, parent guidance
- CODA: AI Companions Safety Guide
- CODA: ChatGPT Safety Settings for Parents
Further Reading
About Children of the Digital Age
Children of the Digital Age provides online safety education and training for schools, parents and organisations. Our guides support practical conversations about online behaviour and the choices people make. However, education and app settings do not replace safeguarding or professional support.
Send a message
Contact Children of the Digital Age to enquire about online safety education, parent sessions or training for your school or organisation.
