AI is rapidly transforming school psychology. Some graduate programs have banned it outright. Others are still figuring out where to draw the line. NASP has convened an AI Task Force and begun developing resources and guidance, but has not yet issued a formal position statement. Their current technology guidelines were not written with large language models in mind, and the gap is evident.
Here is what I think we need to talk about: not whether school psychologists should use AI, but whether we are training them to think well enough to use it responsibly.
The Human Element Is Not Optional
Psychology is a deeply human profession. Any tool we bring into this work needs to serve that, not undermine it. We should be confident that the tools we use are culturally sensitive. We should guard against the illusion of social connection that AI can create. And we need to stop treating AI like it is a professional expert, because it is not one.
That last point matters more than people realize. It is already common for adolescents and young adults to turn to large language model chatbots for mental health advice (McBain RK, Bozick R, Diliberti M, et al. JAMA Network Open, November 2025). If young people are treating AI like a therapist, what happens when school psychologists start treating it like a colleague?
Students Are Already Using It
Let's be honest about the current reality in graduate programs. Students are already using AI, whether or not we have policies that account for it. Most programs prohibit using AI to generate assignments, and that makes sense. But students are using it to summarize complex material, support early-stage research, brainstorm, check grammar, and adjust tone. A student can take an informal case write-up and have AI transform it into something that reads like a professional document in seconds.
None of that is inherently wrong. But it raises a question programs need to answer: Are we teaching students to think, or are we teaching them to produce?
The Critical Thinking Problem
What I see lacking in some programs is explicit instruction in critical and clinical thinking. A 23-year-old NYU graduate recently told me she thinks schools need to bring critical thinking back to education. I think she is right, and I think the problem runs deeper than most of us want to admit.
Teaching classes in experimental design, statistics, and consultation is not the same as teaching students how to use scientific methods to answer questions and solve problems. Those are related skills, but they are not the same skill.
School psychology has always required clinicians to be critical thinkers who use data and scientific reasoning to make decisions and draw conclusions. This is embedded in the NASP Practice Model (2020): Data-Based Decision Making (Domain 1), Research and Evidence-Based Practice (Domain 9). It is implied in Equitable Practices for Diverse Student Populations (Domain 8) and Legal, Ethical and Professional Practice (Domain 10).
These are exactly the skills you need to use AI well. But without explicit training, will students actually apply them? Will they fact-check AI output? Will they even know how to fact-check it?
Next in a Long Line of Tools
School psychologists have always used tools to manage cognitive load. Before computers, there were manuals like the Pre-Referral Intervention Manual (PRIM) to help generate intervention ideas. Then came software like X-BASS, developed by Dr. Ortiz, which analyzes cross-battery assessment data that is challenging and time-consuming to do by hand. Today, school staff searching for ideas are more likely to turn to an AI tool, even if it is just the AI summary that pops up in a Google search. AI is a natural progression.
But the professional expectation has not changed. School psychologists are required to produce individualized reports and IEPs regardless of what tools they use. AI just makes it harder to verify that the expectation is being met. The question is not whether AI belongs in this line of tools. It is whether we are keeping up with what it demands of us.
The Harder Question
Using AI does make our work more complicated in one specific way: it requires us to catch mistakes we may not recognize as mistakes.
What skills do school psychologists need to spot hallucinations, logical inconsistencies, cultural bias, overstatements, critical omissions, or conclusions that do not match the data? When you do not know what you do not know, how do you evaluate what a machine is telling you? There will now be more errors of commission than omission, but a sharp skill set can capitalize on the additional insight while catching errors.
Is this really any different than evaluating credible sources for research, or looking up primary sources to fact-check an expert? In principle, no. In practice, AI makes it easier to skip that step because the output looks polished and confident.
Not long ago, while teaching my class about cognitive dissonance and confirmation bias, I brought up the satirical "Birds Aren't Real" movement. I pointed out how easily people trust the internet, to the point where some still believe the satire is real. A few of my own students paused to consider whether the hoax might actually be true. If trained graduate students hesitate over a parody conspiracy theory, will there be a critical eye when AI delivers an assessment interpretation wrapped in professional language?
Where We Start
When I interview new graduates from school psychology programs, most can answer questions about how they would assess dyslexia, autism, or intellectual disabilities. But many cannot tell me what the competing alternatives to those classifications are. Even experienced school psychologists stumble with that question.
In my experience, this is not usually a lack of training in how to identify an educational disability. Applicants to my schools can tell me how to test for autism, dyslexia, and other conditions. What is often missing is explicit training in critical clinical thinking: the habit of asking, "What else could this be? What am I not seeing?"
When I train my staff, it does not take long for them to start considering the impact of trauma, assessing for anxiety, factoring in executive function deficits, and staying mindful of the bias and limits of our assessment tools with culturally and linguistically diverse students. The capacity is there. It just needs to be activated.
As school psychologists, if we strengthen our ability to think critically, ethically, and scientifically, we do two things at once. We strengthen the supports we provide to children and families. And we prepare ourselves to actually leverage what AI has to offer, instead of being leveraged by it.
