Student Mental Health and AI: Keeping Young People at the Centre
by Dr Gina Cherkowski
Artificial intelligence is rapidly changing how we live, work, learn and connect. For our children, however, the impact may be even more profound. AI is not simply changing the tools they use; it is shaping how they learn, think, communicate, build relationships and experience childhood.
Our young people are growing up in a world where AI is quickly becoming part of their everyday lives. They are using it to learn, create, solve problems, make decisions and, increasingly, to talk about their lives and the things that matter to them. There is tremendous possibility in this technology, but with that possibility comes an important responsibility for the adults who care for and educate young people. As educators and parents, we need to help our young people navigate this changing world thoughtfully and safely, while ensuring that human connection, relationships and well-being remain at the heart of how they grow and learn.
I think often about the young people who may be especially drawn to AI for connection. During my years as a teacher, I recall one 9th grade student who came to me and quietly asked, “Miss C, can you help me make a friend? I want friends.” I have never forgotten that moment.
Sadly, there have been other students over the years who have shared that same sense of loneliness and longing to belong, to be seen and to matter.
I think about those students now as AI becomes something young people can talk to, confide in, and even begin to think of as a friend. For a child who feels lonely or socially isolated, that connection may feel very real and may even provide comfort. We should not dismiss the value of that. But we also need to ask what happens when an AI chatbot begins to take the place of the complicated, sometimes uncomfortable, but deeply important work of building relationships with other people.
The question is no longer whether students will use AI. They are using it. The more important questions are: How will they use it? What might they gain? What might they lose? And how do we ensure AI supports, rather than undermines, the conditions young people need to flourish?
AI is already part of students’ lives
The speed of adoption is remarkable. OECD data indicate that AI chatbots have quickly become part of how many students learn.
Students are also turning to AI for much more personal reasons. A 2026 JAMA Pediatrics study found that 19.2% of U.S. adolescents and young adults ages 12 to 21 had used AI chatbots for mental health advice. More than 90% of those users described the advice as at least somewhat helpful, yet 63% had told no one they were using AI in this way.
That should get our attention.
Young people may be having conversations about anxiety, relationships, loneliness, identity and emotional distress with an AI system without a parent, teacher, counsellor or other caring adult knowing.
At the same time, AI holds real promise. The answer is not to create fear around AI or try to remove it from students’ lives. Instead, we need to help young people develop the knowledge, judgment, social and emotional skills, and relationships necessary to use it well.
Mental health requires greater caution
There has been ongoing discourse online and in schools about student use of AI for learning, and the concerns around outsourcing the learning. It is important that AI does not remove the productive struggle of learning. Students develop persistence by persisting. They develop problem-solving skills by working through problems. And they build confidence by encountering something difficult and discovering, I didn’t know how to do this, but I stayed with it, worked through it, and now I can.
However, there is another kind of outsourcing that has me worried: the outsourcing of human connection to chatbots.
AI can simulate empathy remarkably well. It can respond immediately, patiently and without appearing judgmental. For a young person who is lonely, struggling or simply looking for someone to talk to, that can be incredibly appealing. But simulated empathy is not the same as human connection. A chatbot may sound caring, but it cannot replace the experience of being truly known, understood and cared for by another person.
This distinction is critical: A chatbot that seems caring is not the same as a caring adult who knows a child. It cannot notice the change in a student’s face as they walk into class. It cannot recognize that a student who normally laughs with friends has suddenly become quiet. It does not truly know the young person’s family, culture, relationships or the context of their life.
And there is another important difference: sycophancy. AI systems can sometimes be overly inclined to agree with, affirm or mirror the person they are interacting with, even when challenge, correction or a different perspective may be what that young person needs most.
A caring adult does not simply tell a child what they want to hear. Sometimes caring means asking a difficult question, offering another perspective, setting a boundary or saying, “I hear you, but I see this differently.” This is a natural part of human development. It is part of what makes human relationships so important.
Most importantly, AI cannot take human responsibility for that child.
Protecting belonging, mattering and human connection
This brings us to a much bigger question: How will AI affect the development of our social, emotional and relational skills?
Children and adolescents develop through relationships. They learn to read facial expressions, navigate disagreement, manage frustration, repair relationships, listen, empathize, collaborate, cope with rejection and understand another person’s perspective.
These everyday interactions are where social, emotional and relational intelligence develop.
AI may provide useful opportunities to practise some of these skills. A young person might rehearse a difficult conversation, ask for help understanding another perspective or practise what they want to say before approaching a teacher, parent or friend.
But practicing a relationship is not the same as being in one.
Human relationships require reciprocity. They ask something of us. We have to notice how our words affect another person, tolerate misunderstanding, compromise, apologize, forgive, listen and sometimes sit with discomfort. These are not inconveniences we should engineer out of childhood. They are part of how we become socially and emotionally capable human beings.
This is why social and emotional learning is becoming more important in the age of AI, not less. SEL is mental health promotion. Self-awareness, emotion regulation, empathy, responsible decision-making, relationship skills and help-seeking provide young people with the capacities they need to navigate both their human and digital worlds.
They also help create something every child needs: a sense of belonging and mattering.
Young people need to know, I belong here. I matter. People see me. People know me. My presence matters. Someone would notice if I wasn’t here.
For a young person experiencing loneliness, those experiences of authentic connection may be particularly important. AI may offer companionship in a moment, but our larger responsibility is to ensure that young people have opportunities to build meaningful relationships with people.
Protecting what it means to be human
As machines become better at writing, creating, answering questions and even simulating empathy, distinctly human capacities such as curiosity, compassion, judgment, creativity, courage, connection and care become more important, not less.
UNESCO’s AI competency frameworks place a human-centred mindset at the foundation of AI literacy, alongside ethics, AI knowledge and responsible creation.
For schools and families, this means intentionally protecting the experiences through which human development occurs.
Children still need conversation around a dinner table. They need recess, play, sport, music, art and time outdoors. They need friendships that sometimes become complicated. They need adults who listen to them. They need opportunities to contribute, belong and know that they matter.
Perhaps one of our simplest principles should be less screen and more green.
This does not mean rejecting technology. It means creating balance.
Our children need to understand AI, and they need opportunities to explore the natural world. They need digital literacy and relational literacy. They need technological fluency and human connection. They need to know how to work with intelligent machines and how to sit across from another human being, look them in the eye, listen, empathize and care.
We need both to thrive in today’s complex world.
I caution however that AI literacy cannot simply mean knowing how to write a better prompt. It must include knowing when not to use AI, when to question it, when and how to verify information, when to protect private information and, perhaps most importantly, when to put the technology down and turn toward another human being and become an active and engaged participant in the world around us.
What should schools and families do now?
There are some practical principles that can guide us.
Protect human connection. AI should complement, not substitute for, relationships with teachers, parents, counsellors, coaches, friends and trusted adults.
Teach AI literacy alongside social and emotional literacy. Students need to understand bias, hallucinations, privacy and misinformation, but they also need self-awareness, boundaries, empathy, help-seeking and healthy relationship skills.
Keep productive struggle in learning. Before allowing AI to complete a task, ask what thinking the student needs to learn and practice.
Create clear boundaries around mental health. AI may provide general information or help a young person prepare questions for a professional, but general-purpose AI and companion chatbots should not replace appropriate human mental health support.
Protect time for human experiences. Technology should not crowd out conversation, play, movement, nature, creativity and relationships.
Listen to young people. Ask students how they use AI, what they value about it, what makes them uncomfortable and what they do when an AI response does not seem right.
And above all, keep humans in the loop and at the centre.
The question is not AI or Humanity
I am optimistic about what AI can contribute to education. It can help personalize learning, remove barriers, increase accessibility, spark creativity and potentially give educators more time for the deeply human parts of their work.
But we should be equally clear about what we are trying to protect.
The future our children are entering will require technological fluency. It will also require deeply human capacities: a strong sense of identity, belonging and mattering; emotional regulation; empathy; courage; curiosity and imagination; critical thinking; healthy relationships; and the ability to make thoughtful decisions and persist when an answer is not immediately available.
Our responsibility is not simply to prepare children to use AI. It is to prepare them to remain deeply human in a world increasingly shaped by AI.
That means teaching young people not only how to ask AI better questions, but how to question AI itself. Not only how to get an answer faster, but when the struggle to find an answer matters. And not only how to connect with increasingly human-like technology, but how to continue choosing, building and investing in one another. AI should expand human potential, not replace the experiences and relationships through which that potential develops.
We can be optimistic about AI while being thoughtful about what we allow it to replace, but we need a balance.
Technology and humanity. Innovation and connection. Less screen, and more green.
References
Feng, X., et al. (2025). The Effectiveness of AI Chatbots in Alleviating Mental Distress and Promoting Health Behaviors Among Adolescents and Young Adults: Systematic Review and Meta-Analysis. Journal of Medical Internet Research, 27.
McBain, R. K., Cantor, J. H., Breslau, J., et al. (2026). AI Chatbot Use and Disclosure for Mental Health Among US Adolescents and Young Adults. JAMA Pediatrics, 180(8), 884–890.
OECD. (2026). OECD Digital Education Outlook 2026: Exploring Effective Uses of Generative AI in Education. OECD Publishing.
Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi.org/10.1016/j.tics.2016.07.002
UNESCO. (2024). AI Competency Framework for Students.
Winthrop, R. (2025). Why We Need a Minimalist Mindset When It Comes to AI and Tech Use for Young People. Brookings Institution.
Dr. Gina Cherkowski
Dr. Gina Cherkowski is an educator, researcher, award-winning social entrepreneur, and a national leader in school mental health, social and emotional learning (SEL), belonging, and educational well-being. She is the Executive Director of the Werklund Institute for Social and Emotional Intelligence Canada Ltd., where she leads national initiatives focused on strengthening mental health and well-being, SEL, relational intelligence creating healthy, thriving schools and systems.
Dr. Cherkowski’s work focuses on translating research into practice and creating the conditions that allow students, educators, and leaders to thrive. An International CASEL Fellow and two-time TEDx speaker, she has worked around the globe and across education, research, and community sectors to advance evidence-informed approaches to learning in a complex world.
Her current work explores the increasingly important intersection of AI, mental health, and human skills, with a focus on ensuring that as technology evolves, education continues to intentionally strengthen empathy, critical thinking, creativity, relationships, emotional intelligence, and the sense of belonging that make us uniquely human.