The Hidden Side of Student AI Use: Mental Health, Access and Risk

Why AI Fills a Gap

Why might a student confide in a chatbot when educators, family members and friends are available? Perhaps because AI is accessible at any hour, costs little or nothing, responds immediately and feels nonjudgmental. For a student worried about embarrassment, stigma, family reaction or whether an adult will keep a disclosure private, using a chatbot may seem like a low-risk first step. 

Access also matters. A chatbot may help a student in various ways, whether assisting with schoolwork, helping process feelings, preparing them to approach an adult, or avoiding feelings of loneliness or depression. But a chatbot’s response does not equal care or understanding.

As Buddy Berry, superintendent of Eminence Independent Schools in Kentucky, explains, “AI can be a powerful tool for students, but it can never replace the meaningful relationships and trusted adults that are essential to student well-being.” A system can simulate compassion, but it cannot understand the full context or take responsibility for continuing support.

A chatbot’s ability to sustain a conversation can make the interaction feel human. However, conversational fluency is not the same as understanding, professional judgment or genuine human connection. A system cannot fully understand a student’s circumstances or take responsibility for the continuing support that a student may need.

DISCOVER: Here’s how schools can recognize and stop AI-generated bullying tactics.

Risks Schools Cannot Ignore

Educators need to help students understand how these systems work and how to stay safe while using them. A chatbot may miss warning signs, reinforce a student’s thoughts, offer inappropriate advice or continue a conversation that should instead be handled by a qualified professional. It may sound empathetic, but it lacks the clinical judgment needed to recognize when a student needs immediate human support. 

Privacy is another concern. Students may disclose details about trauma, self-harm, family conflict or health without realizing that conversations may be retained, reviewed, used or shared under a product’s terms. Nancye Blair Black, AI in education expert consultant and founding CEO of The Block Uncarved, says, “Most consumer and educational AI technologies are not protected by the HIPAA, meaning that when students share intimate details about their personal life, physical health or mental health status, this data may be used to train future AI models, sold to other companies for unforeseen purposes, or even used to inform targeted ads or messaging that might exploit the child's vulnerabilities and threaten their agency.”

An additional concern is bias, which can make the experience less safe when students’ language, culture, identity or disability is not adequately represented. The American Psychological Association advises against the reliance on generative AI chatbots as a replacement for mental health treatment and calls for safeguards. District AI guidance should include language explaining that chatbots may simulate empathy but should not be relied upon in place of a counselor, crisis service or trusted adult.

Click the banner below to keep receive the latest K-12 EdTech content directly in you inbox.

Building a District Response

For K–12 districts prioritizing safe AI use, the first step is to determine how students are using these tools. In my work, I have used anonymous surveys, led student focus groups and held panel discussions that help reveal why students turn to AI and what prevents them from seeking human support. Involving families in the conversation can help districts understand what they are seeing when students are not in school.

Kristen Mattson, a lecturer at the University of Illinois, acknowledges that learning how students interact with these tools “can feel scary” and activate educators’ natural desire to protect them. However, she cautions that “just saying ‘no’ glosses over the root issues causing young people to turn to AI in the first place.” Instead, she encourages educators to “be curious about the lived experiences of our students, invite them into judgement-free dialogue, and let them know they have an ally in us.” She also recommends educating families and community members about the appeal and risks of AI companionship.

UP NEXT: Discover how universities and colleges are supporting K–12 AI adoption.

It is essential to understand what leads students to turn to AI rather than seek human support. Students are asking for guidance. Jessica Gillespie, director of instructional improvement–innovation at Griswold Public Schools in Connecticut, says students “want guidance around AI, not just to be told not to use it.” District frameworks should define boundaries for emotional support use, identify personal information students should never enter and explain when a student should stop an AI conversation and contact a human for support.

Districts also need rigorous reviews of AI-enabled wellness products, including age requirements, evidence of effectiveness, accessibility, bias, data practices, human oversight and crisis escalation procedures. A supportive tone alone does not make a product “wellness technology.” Gillespie shared, “Our district work on digital wellness includes work to incorporate lessons that explicitly teach students to consider their online behavior and to understand how AI works, so that they understand the risks that can come with online behaviors.” 

Most important, schools must strengthen the human network around technology. Staff need training to recognize concerns and make referrals. Students should have several adults they can approach. Partnerships with community agencies, telemental health experts and families can expand support when school personnel are stretched thin.

AI literacy and student wellness should not be separate initiatives. Students can practice evaluating emotionally persuasive responses, protecting sensitive information, recognizing when a chatbot is affirming rather than helping and moving from an AI conversation to human support.

The main question is not, “How do we stop students from using AI this way?” It is, “What need is this use revealing, and how do we make safe human support easier to reach?” Districts that ask that question will create AI policies that do more than control technology. They will build stronger systems of safety and human connection around students. Berry explains, “Our responsibility is to teach students to use AI safely and responsibly while ensuring every student knows there are real people in their school who are ready to listen, support and help them when they need it most.”