|
By the time a kid walks into your office at 10 a.m., there is a real chance they have already talked to an AI chatbot about their anxiety, their parents, or whether life is worth living. AI mental health apps in schools are not a future problem. They are a present caseload variable. If you are a school counselor, school social worker, school psychologist, BCBA, or LMFT working in K-12, this post is a plain-language read on what is happening, why it matters, and what to do about it on Monday.
In May, Common Sense Media and Stanford University's Brainstorm Lab released a risk assessment of the most popular AI mental health apps teens are using. The headline finding: the direct-to-consumer market is largely unregulated, the products vary wildly in quality, and several of them produce responses that range from useless to actively harmful. Apps designed for use inside schools, with humans kept in the loop, performed meaningfully better. EdWeek covered the assessment the same week.
Why AI Mental Health Apps in Schools Are Suddenly Everywhere
A few things converged at once. First, post-pandemic demand has not come down. The Bureau of Labor Statistics still projects 8 to 10 percent growth in school counseling and educational guidance roles through 2032, and most of us already feel underwater. Kids are turning to whatever is available between sessions, and what is available is their phone.
Second, federal school mental health funding has been on a rollercoaster. The Trump administration pulled roughly $2 billion in mental health grants in early 2026 and restored them about a day later. The Department of Education has since awarded $208 million in new mental health grants, but that is less than what was paused earlier. Congress set aside at least $164 million for the two main school mental health grant programs in the fiscal 2026 budget. The net effect: districts are nervous about long-term staffing, and tech vendors are pitching AI tools as the gap-filler.
Third, the apps have gotten genuinely sophisticated. A 14-year-old can have a conversation that feels supportive at midnight, when nobody on your team is on call. That is not nothing. It is also not therapy, and the distinction matters more than the marketing copy suggests.
What the Common Sense and Stanford Assessment Actually Found
The assessment looked at popular AI mental health apps across two categories: direct-to-consumer products that any teen can download, and school-deployed products that come through a district or vendor relationship.
Three findings stood out to me.
Direct-to-consumer apps are not bound by clinical standards
Most of the apps teens are downloading were never designed as clinical tools. They are wellness products. They are not required to follow crisis protocols, mandatory reporting standards, or any of the practice frameworks the rest of us are licensed under. Some of them encouraged unsafe behavior or failed to escalate clear suicide risk language.
Human in the loop makes a measurable difference
Products designed for school deployment that include a real clinician reviewing flagged conversations, intervening when risk is detected, or shaping the model's responses performed substantially better. This tracks with everything we already know about telehealth and digital mental health: the technology is a delivery vehicle, not a clinician. (ASCA, NASW, and NASP have all said versions of this in their digital practice guidance.)
Kids will not tell you which apps they use
In the focus groups embedded in the assessment, teens described their AI conversations the way a previous generation described journaling. Private. Personal. Not something they planned to bring up unless asked directly. That changes how you do intake.
What This Means for School-Based Practice
You do not need a position statement to act on this. Here is what I would do this week.
Add one screening question
Pick whatever your standard intake or check-in looks like, and add one question: "Do you ever use AI apps or chatbots to talk through things that are bothering you?" Follow up if yes. Which one. How often. Has it ever told you something that did not feel right. Has it ever told you to do something. You will be surprised by what you learn.
Get clear on what your district is paying for
If your district has signed on with an AI-assisted mental health platform (Daybreak, BeMe, Wysa for Education, and several others are active in the school space), you are part of the human-in-the-loop. Find out: who reviews flagged messages, what the response time SLA is, what happens after hours, what the mandated reporting workflow looks like. If nobody can answer those questions, that is your finding.
Update your safety planning language
Most of our safety plans were written assuming kids reach out to a person. Build in a line about what to do if an AI tool gives advice that feels off, encourages something unsafe, or is the only thing they have used during a crisis. Teach the kid that those conversations are evidence you want to hear about.
Document what you know and what you do not
If a student reports that a chatbot told them to do something concerning, that is reportable to the same channels you already use. If a parent asks what apps are safe, your honest answer is that the field does not yet have a consensus list. Saying so is not a weakness. It is accurate practice.
The Profession-Specific Angles
A few notes for specific roles, because the picture is not identical across disciplines.
School counselors (ASCA-aligned, ACA-licensed)
ASCA's digital and social media guidance already covers a lot of this terrain. The newer wrinkle is that the ACA Code of Ethics, particularly Section H on distance counseling, applies even when the "distance" is a tool a student is using outside your session. Document AI use in the same place you document any other clinical variable.
School social workers (NASW, LCSW)
NASW's Technology Standards still hold, and the supervision question is the one to watch. If you carry an LCSW and supervise interns or LMSWs working with students who are using AI tools heavily, build that into your supervision agenda. CE programs are starting to cover this; expect more clock hours focused on AI literacy over the next year.
School psychologists (NASP, NCSP)
NASP's 2026 PPI in July is going to lean hard on advocacy, and digital mental health is on the agenda. If you carry the NCSP, this is exactly the kind of issue that belongs in your continuing education plan.
BCBAs working in schools
This is its own animal. AI tools rarely show up in ABA-specific work the same way, but they show up in the home environments, the peer relationships, and the emotion regulation contexts that interact with your programs. The BACB's January 2026 transition off the ABAI VCS to the new coursework attestation system has occupied a lot of bandwidth, but ACE-eligible content on the intersection of digital tools and behavioral programming is starting to appear. EduCare's BACB ACE provider status (OP-26-12340) covers content in this lane.
LMFTs and AAMFT-affiliated practitioners
Family systems work picks up a new variable: the kid is using an app the parents do not know about, and the app is shaping how the kid talks about the family. That is a clinical observation worth naming in session.
A Few Things I Am Not Saying
I am not saying these apps are uniformly bad. The school-deployed, human-in-the-loop products are doing real work in districts that genuinely cannot hire enough clinicians, and that includes districts in California, Texas, South Carolina, and several other states scaling these contracts right now. I am not saying we should be the AI police. Banning the conversation will not make the conversation stop. It will just stop the kid from telling you. I am saying: this is a fast-moving piece of the clinical landscape, your training did not cover it, and the kids using these tools deserve a practitioner who knows the basics.
Where to Go Next
If you want to go deeper, the original Common Sense and Stanford Brainstorm Lab assessment is worth an hour. EdWeek's coverage gives you the policy frame. NASP, ASCA, and NASW are all publishing in this space, and if you carry the BACB credential, expect ACE-eligible content on related topics to expand through 2026.
EduCare runs continuing education in school-based practice, clinical tools, and law and ethics across the disciplines named in this post. Our BACB ACE provider status (OP-26-12340) is approved. Our NBCC ACEP accreditation is pending, and once active, courses for counselors will reference clock hours accordingly. If you want a course list aligned to what you actually see in school buildings, that is what we build.
The kids using these tools tonight will be on your caseload tomorrow. Knowing what to ask is the first move.
| 
