To put it simply, kids are not okay. Middle and high school students nationwide carry an invisible backpack to school—filled with stress, trauma from past experiences, and anxiety. The CDC reports that teen rates of anxiety, depression, and suicidal thoughts have skyrocketed 40% over the past 10 years. A whopping 69% of schools report an increase in students seeking mental health services, and only 13% can meet the need, according to the Condition of Education report from NCES.
As the student mental health crisis continues to grow, schools struggle to find school-based mental health professionals to meet students’ needs. Our country would need to add 100,000 school-based mental health professionals to provide adequate support throughout the public school system. Traditionally, students have received mental health support from school-based clinicians; however, it would take years to train enough school staff to meet the needs of middle and high school students. Because schools won’t be able to fully staff counseling positions soon, we must consider new ways to support student well-being and flag any serious concerns before they become critical.
Advances in technology now make it possible for schools to deliver prevention and early intervention to all students and triage those who might need more intensive services. By tapping into generative AI tools, schools can provide personalized, timely mental health interventions that can improve student well-being. While AI will never replace human counselors, it can help bridge a critical mental health gap.
Technology Applications in Mental Health
When educators, parents, and community members see the term “AI” used with student mental health, they often question these tools’ privacy, efficiency, and effectiveness. Many people are unaware that experimentation with chatbots in psychology has been underway for decades. This research reveals that, in addition to making mental health support more affordable and accessible, chatbots are often a preferred entry point to therapy for people who are skeptical or have social anxiety. This can be especially true for teens, who are often uncomfortable opening up to adults.
Computer scientist Joseph Weizenbaum developed the first AI chatbot for mental health in 1967, named Eliza. It simulated a psychotherapist in natural language conversations with users. Eliza worked by scanning user inputs for keywords and responding with pre-written responses based on those keywords. If it couldn’t match user inputs to keywords, it gave generic prompts like “Tell me more.” In his work, Weizenbaum noticed that people talking to Eliza attributed understanding, empathy, and other human characteristics to the software.
Today’s generative AI continues to be only as good as the data it uses as its source of knowledge. However, when we supplement or augment its knowledge with clinical knowledge and carefully crafted guardrails, we can effectively harness the power of AI to provide personalized coaching and skill building. Take a moment and think about this for yourself…if you were having a conflict with a friend or loved one, would you rather someone hand you a worksheet on how to use an “I statement” to express how their actions hurt you or would you prefer a conversation where you get validation for how frustrating it is and after talking it through you gently realize hmm… I should maybe just tell them how I feel using an I statement. As this example demonstrates, AI can facilitate teaching a skill (I-statements) in a way that feels validating, empowering, and accessible, as you may not feel comfortable talking about certain topics with a human as a first step. In fact, in our research at Alongside, we found that .02% of students reached out to a counselor via the app on their own about what they were struggling with; however, after chatting with our clinician- powered AI chatbot, 37% of students shared what they chatted about with their school counselor. This suggests that AI-facilitated conversations may increase comfort around the topic to the point where students feel more comfortable reaching out for in-person support.
Leverage AI to Reach the Unreached
My role at Alongside is to consistently tackle how to enact our mission of providing evidence- based, personalized, preventative mental health support using technology. We, therefore, have evolved to embrace generative AI as the technology has rapidly advanced in the past two years. We have embraced generative AI as it allows us to provide personalization and convey empathy and validation in ways other types of technology cannot. Our core feature of Alongside is an AI- powered chatbot that helps students learn and apply evidence-based problem-solving, social skills, mindfulness, and emotional coping skills to their lives and unique personal experiences. Written by experienced clinicians, the AI chatbot follows carefully crafted scripts that guide the students through learning and applying skills to leave them feeling motivated, empowered, and heard. In addition to guiding students, we also monitor to ensure students who need more support are identified and connected with the school counseling team.
Here’s how it works. Students log into the chatbot with their school email and can chat, track their mood, journal, or explore resources. They can discuss any topic and get matched with a relevant and personalized clinician- developed chat exercise. The chatbot asks students to set a goal that focuses on taking one small step in real life to increase hope and self-efficacy, critical components of resilience. Students receive encouraging reminders to reinforce following through on goals they set in the app.
AI chatbots are an exciting new tool that can provide Tier 1 support personalized to the student and requires minimal teacher or counselor effort to implement. With the overwhelming number of students who need support for AI, it can support extending the reach of counselors and prevent issues from escalating to become crises.
Over time, this can result in caseload reductions for school mental health teams while cost- effectively meeting every student’s needs. These apps can also provide schools with the data to better understand the issues most affecting students.
Identify Severe Issues More Quickly
Beyond allowing students to vent about their feelings, AI chatbots can help school leaders monitor students struggling with more severe issues, such as self-harm, suicidal ideation, homicidal ideation, and abuse. Our research shows that the school teams did not previously identify most students who discussed severe issues in the chatbot as struggling students. When students using the Alongside app did bring up thoughts of harming themselves or others, data shows that school teams were immediately alerted and could then guide them through a safety protocol. Notably, the small percentage of students who disclosed thoughts of self-harm during a chat session tended to be high-risk according to the Columbia Suicide Severity Rating Scale (CSSRS). This emphasizes how this intervention can be an on-ramp to getting more intensive, in-person help when needed.
Students Prefer Confidential, Personalized AI Solutions
Today’s tech-savvy teens prefer communicating in a way that’s familiar to them. In 2023-24, Alongside partnered with middle and high schools across the United States, giving them access to our AI chatbot. Our findings from over 30,000 chat sessions overwhelmingly demonstrated that teens value a safe space to explore and process the world around them without judgment.
Over 36% of the time, students using the app chose to engage in chats to process their feelings and feel “heard.” Students highlighted that they didn’t always want or need advice; they felt that receiving empathy and encouragement helped them overcome a challenge.
- The most common issues causing teen stress are related to navigating peer and family relationships. Over 40% of the time, teens discussed or sought support for an interpersonal relationship.
- Teens consistently reported that getting help in a private, confidential space gave them a way to talk about issues they did not yet feel comfortable speaking to an adult about. Specifically, teens noted they didn’t have other opportunities to engage in confidential conversations or were worried about being judged, making them hesitant to reach out.
- Contrary to common conceptions and what we see in outpatient services, boys seek support when given access to a platform that meets them where they are. On average, male, female, and non-binary students spent roughly the same time on the app.
- The app gave students the confidence to seek outside help if they felt they needed it. Thirty-seven percent of students who completed a chat reported sharing a summary of their chatbot conversations and goals with their school counselor.
A Tool, Not a Replacement
When I talk to schools and school counselors, the biggest concern I hear about AI-powered student mental health solutions is that the technology will ultimately replace them. Instead, educators should view the app as an extra layer of support that helps students connect to trained counselors or staff members. By providing a confidential space for students to access help, schools using Alongside reach students who would otherwise not seek help. In most districts, this equates to an additional 25-35% of students receiving individualized support regularly, with students choosing to share information about the support they receive in the app with their counselor about 37% of the time. Beyond this, Alongside typically identifies severe issues that would otherwise go undetected in about 2% of students.
Of course, school leaders must be sensitive to privacy concerns and understand the limitations of what an AI-powered app can deliver. Understandably, school leaders and parents are more apt to trust AI solutions with tight safeguards around students’ personal information and private chats. For example, only when a student mentions harming themselves or others will Alongside send a notification to the student’s school. It also complies with the Children Online Protection Privacy Act (“COPPA”) and the Federal Education Rights and Privacy Act (“FERPA”), which allows school leaders or students’ legal guardians to access their data and request its deletion.
Generative AI tools developed by clinicians and backed by research hold promise for improving adolescent student well-being. By allowing middle- and high school students to engage in personalized conversations that address their anxiety, depression, relationship challenges, and more, schools can effectively bridge the mental health gap, providing equitable support for every child.
About the author
About the author: Dr. Elsa Friis, Ph.D., MSc-GH, is a licensed psychologist, researcher, and behavioral health advocate who specializes in utilizing technology to expand access to youth mental health support. She enjoys leveraging her clinical, research, and public health training to ensure schools can deliver evidence-informed support to students nationwide.
