What Schools Are Missing On About AI: Authority, Evidence & Isolation

Education World Wide’s Lidija Elezovic offers this commentary on what schools are missing about AI. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

As someone who works daily with Gen Z and Gen Alpha students in the role of both school counselor and psychologist, I have watched the conversation around AI in schools become increasingly polarized. Some people see AI as an academic shortcut that will destroy learning. Others see it as an inevitable tool that schools simply need to accept and integrate. In my experience, both views are incomplete.

AI is already being used by students in many different ways. Some use it to complete assignments. Some use it to organize their thoughts. Some use it to study, summarize material, ask questions they are embarrassed to ask in class, or rehearse difficult conversations. Others use it for emotional reassurance, identity exploration, or companionship.

That range of use matters. When adults talk about AI in schools, they often focus almost exclusively on cheating or academic integrity. Those are valid concerns, but they are not the only concerns. From a psychological and developmental perspective, the more important issue is not only whether students are using AI. It is how they are relating to it.

The students I work with are growing up in a world where answers are immediate, personalized, and delivered with a level of confidence that can feel authoritative. For young people whose cognitive, emotional, and social development is still in progress, that creates new challenges that schools, parents, and mental health professionals need to understand.

One of the first things I think people get wrong is the assumption that AI is simply a neutral tool. It can be used as a tool, of course. It can help students brainstorm, review material, simplify difficult concepts, or practice skills. But many young users do not experience AI as just a tool. They experience it as a responsive presence that answers instantly, adapts to them, and often sounds certain.

That is where I am beginning to observe several developing trends in how younger generations are using AI.

Authority Inflation

The first trend is what I refer to as authority inflation.

Because AI-generated responses are immediate, personalized, and presented in confident language, younger users may begin to assign them more credibility than they deserve. In some cases, students may view AI responses as more reliable than feedback from teachers, parents, counselors, or peers.

This is not necessarily because students are intentionally rejecting adults. It is often because AI feels easier. It does not appear tired, impatient, disappointed, or judgmental. It responds quickly. It gives an answer that sounds complete. It can validate the user’s question and produce a polished explanation within seconds.

For students, especially adolescents, that can be very persuasive.

Adolescence is already a stage where young people are negotiating independence, authority, identity, and belonging. They are beginning to question adults, test ideas, and form their own worldview. That is normal development. However, when a student is repeatedly turning to AI for explanations, reassurance, or interpretations of reality, the confidence of the response can begin to matter more than the quality of the evidence behind it.

This is what concerns me. A confident AI response can appear more credible than a nuanced human response. A teacher may say, “It depends,” or “Let’s look at the evidence,” or “That is more complicated than it seems.” AI may provide a clear, organized answer immediately. To a young person, clarity can be mistaken for truth.

In most students, this may simply lead to overreliance or poor critical thinking. But in vulnerable students, authority inflation may weaken reality testing. If a young person is already struggling with anxiety, paranoia, obsessive thinking, or distorted beliefs, an AI response that reinforces their fears or gives structure to a speculative thought can become problematic.

This is one reason I think we need to be careful with the phrase “AI psychosis.” AI itself does not create psychosis. Technology cannot independently produce a psychotic disorder in someone who has no underlying vulnerability. However, it can exacerbate pre-existing vulnerabilities. It can intensify certain thought patterns. It can provide repeated reassurance in ways that feed obsessive loops. It can validate fears that should be gently questioned by a trained adult.

That distinction is important. We should not create moral panic. But we also should not ignore the fact that some students are interacting with AI in ways that affect how they understand reality, authority, and evidence.

Confusing Imagination With Evidence

The second trend I am observing is a growing confusion between imagination and evidence.

Adolescents naturally spend a great deal of time exploring who they are. They think about identity, purpose, relationships, spirituality, morality, social status, and the future. This is not new. What is new is that they now have access to tools that can generate elaborate narratives around these topics instantly.

A student can ask AI why they feel different from others. They can ask what their dreams mean. They can ask whether a relationship is toxic. They can ask whether a certain event is a sign of something larger. They can ask questions about personality, destiny, spirituality, trauma, or mental health. In response, AI can produce language that feels deeply personal, meaningful, and convincing.

The problem is that meaningful does not always mean true.

AI can create stories. It can identify patterns. It can present possibilities. But students may not always understand the difference between a speculative interpretation and an evidence-based conclusion. This is especially true when the response is written in emotionally resonant language.

For example, a student who is anxious may ask AI to interpret a friend’s behavior. The AI might provide several possible explanations, but if the student is already fearful of rejection, they may focus only on the explanation that confirms their fear. A student struggling with obsessive thinking may repeatedly ask similar questions until they receive an answer that feels certain enough to temporarily reduce anxiety. A student exploring identity may accept an AI-generated description as a fixed truth about who they are, rather than one possible reflection among many.

This does not mean students should never use AI to explore ideas. Exploration can be healthy. But it becomes risky when students do not have the maturity, support, or critical thinking skills to distinguish imagination from evidence.

Schools often teach students how to evaluate sources for research papers. That skill now needs to extend to AI-generated responses. Students need to learn that AI can sound informed even when it is speculating. They need to understand that a personalized answer is not the same as a clinically informed answer, a spiritually authoritative answer, or a factual answer.

This is where AI literacy becomes much more than a technical skill. It is also a psychological skill. Students need to ask: Is this response based on evidence? Is it making assumptions about me? Is it confirming what I already fear? Is it helping me think more clearly, or is it making me more certain about something I have not actually verified?

Those questions are essential in an AI-enabled classroom and in an AI-enabled adolescence.

Social Substitution

The third trend is social substitution.

Many adolescents report using AI for extended periods of time because interacting with AI feels easier than interacting with other people. This is not difficult to understand. Human relationships are complicated. Friends disagree. Parents set limits. Teachers challenge students. Counselors ask difficult questions. Peers misunderstand, reject, tease, or disappoint each other.

AI, by contrast, can feel smooth. It is available. It responds. It can be supportive on demand. It can be asked the same question repeatedly. It does not require the student to manage another person’s emotions, boundaries, facial expressions, tone, or expectations.

For a student who feels lonely, socially anxious, or misunderstood, that can feel like relief.

However, normal psychological development requires human relationship. Young people need disagreement, unpredictability, frustration, repair, compromise, and social feedback. They need to experience what happens when someone sees things differently. They need to learn how to tolerate awkwardness, apologize, clarify, disagree respectfully, and remain connected even when a conversation is not perfectly validating.

AI cannot fully provide that. It may simulate conversation, but it does not participate in relationship the way another human being does. It does not have its own needs. It does not create the same mutuality. It does not truly require the student to practice empathy, patience, or repair.

That does not make AI useless. It can be valuable as a resource tool. It can help students practice language for a conversation, organize emotions before speaking to someone, or review coping strategies. But it becomes concerning when it replaces human interaction rather than supporting it.

The isolated youth are the students I worry about most.

My biggest concern is not necessarily the student who occasionally uses AI too much for homework. My biggest concern is the student who is already isolated, sleeping poorly, struggling with anxiety or depression, and spending long periods of time in AI interactions that feel safer than real life. When excessive AI use combines with isolation, poor sleep habits, anxiety, paranoia, obsessive thinking, or other mental health challenges, the risk for distorted thinking increases.

In those cases, AI may not be the original cause of the problem, but it can become part of the pattern that maintains it.

What This Means for Schools

This is why schools need to move beyond both fear and hype.

If schools focus only on academic integrity, they will miss the social and emotional implications of AI use. If they focus only on innovation, they may overlook the developmental risks. The reality is that AI can be helpful and harmful depending on the student, the context, the purpose, and the level of adult guidance.

Teachers are not becoming irrelevant. If anything, their role is becoming more important. Students need adults who can help them question information, tolerate uncertainty, and understand the difference between support and substitution. They need educators who can design assignments that value process, not only final answers. They need counselors and parents who are willing to ask not just, “Are you using AI?” but, “What are you using it for?” and “How do you feel after using it?”

AI literacy should be taught alongside traditional skills because students now need to understand not only how to use AI, but how to interpret it. They need to know when AI is appropriate for brainstorming or study support, and when it is interfering with learning, emotional regulation, or relationships.

They also need clear guidance around deepfakes, privacy, and digital consent. AI-generated images, audio, or messages can quickly become tools for bullying, humiliation, or manipulation. Students must understand that creating or sharing manipulated content of another person is not harmless. It can cause real psychological and social damage.

The broader issue is trust. Students need to learn how to trust evidence over confidence, human relationships over simulated validation, and their own developing judgment over instant answers.

A Balanced Response

I do not believe AI should be treated as an enemy of education. I also do not believe it should be treated as a harmless shortcut or an inevitable replacement for human instruction. The most responsible position is somewhere in the middle.

AI can support learning. It can help students access information, organize thoughts, and practice skills. But it should not become an unquestioned authority. It should not replace the thinking process. It should not become a substitute for teachers, counselors, parents, or peers.

What people often get wrong about AI in schools is that they frame the issue too narrowly. They ask whether students will cheat. They ask whether teachers will be replaced. They ask whether AI should be banned or embraced.

Those questions matter, but they are not enough.

The deeper question is how AI is shaping the way young people understand knowledge, authority, evidence, relationships, and themselves. From what I am observing, that is where schools need to pay closer attention.

The students who will thrive in an AI-enabled world will not simply be the ones who know how to prompt a chatbot. They will be the ones who can use AI critically without surrendering their judgment to it. They will be the ones who can accept support without becoming dependent on it. They will be the ones who can recognize that instant answers are not always wisdom, and that human connection, even when imperfect, remains essential to healthy development.

AI may be new, but the developmental needs of students are not. They still need guidance, boundaries, critical thinking, meaningful relationships, and trusted adults who can help them make sense of the world.

That is the part we cannot outsource.