Your Students Know AI Is Weakening Their Thinking. But They Are Using It Anyway.

You have students in your building right now who are using AI even though they believe it is weakening their thinking.

They are not confused about what is happening. Many of them have already noticed the shift.

And they are continuing to use it anyway.

Recent survey data asked thousands of middle school, high school, and college students about their AI use. Nearly seventy percent of middle and high school students said they believe AI is undermining their critical thinking. Among middle school students, that number rose sharply in a matter of months.

At the same time, usage increased. Students are not pulling back. They are relying on it more.

Students are naming the impact while increasing the behavior.

Students are continuing to depend on AI even when they believe it is weakening their ability to think.

In classrooms, this shows up when a student opens an AI tool before attempting a task she already knows how to start. She reads the response, adjusts her work to match it, and gradually stops initiating thinking on her own. Nothing appears wrong. The work is complete. The language is strong.

In math classes, a student who understands the process still checks AI for confirmation or speed. Over time, he begins to rely on it first. When he encounters a problem without that support, he hesitates longer than he used to.

In counseling offices, students describe needing help to “get started” or wanting to “make sure it sounds right” before they begin. They may not name AI directly, but the pattern is there.

In college and career planning, students accept AI-generated recommendations without testing them against their own preferences. The decision feels finished before they have worked through it themselves.

These are not isolated moments. They are patterns forming quietly across settings.

The moment students feel stuck is where dependence begins, and in most cases, that moment is happening outside of adult visibility.

This is not a question of rule-following. It is a shift in how students relate to thinking.

Students begin to move around struggle instead of through it. They lose opportunities to build stamina, persistence, and confidence in their own thinking. Over time, they may stop trusting themselves to begin without assistance.

Instruction becomes harder to interpret because completed work no longer reflects how much support was needed to produce it.

This also creates uneven experiences. Students who rely heavily on AI may appear more proficient than peers who are still working through problems independently, even when their understanding is similar.

The concern is not just about academic performance. It is about whether students are continuing to build the habits that allow them to think without assistance when it matters.

What this looks like to address in real school settings

Schools can respond to this without removing AI access, but they do need to account for how students are already using it outside of school.

In classrooms where AI is part of instruction, teachers can require a visible first attempt before any tool is used. At the start of a task, students write a sentence, sketch a diagram, outline their thinking, or attempt the first step of a problem during class time. They are asked to show where they are unsure or stuck. This creates a baseline of their own thinking before AI enters the process.

If students use AI after that point, they explain what the tool contributed, what they accepted or rejected, and how it changed their approach. This can happen through a short written reflection or a quick conversation during class. The focus stays on their decisions, not the tool’s output.

In classrooms where AI use is restricted, teachers can still build the same visibility. Students begin work in class, talk through their reasoning with a partner, or annotate their process alongside their final answer. The goal is not to monitor compliance. It is to make thinking observable before the work leaves the room.

Teachers can also add a short reflection at the end of assignments where students explain how they arrived at their final answer. Not what the answer is, but how they got there, where they struggled, and what helped them move forward. This keeps the learning process visible even when the final product looks complete.

Instructional leaders can support this by helping teams identify where these moments belong in existing lessons. This does not require new systems or new units. It requires small, consistent checkpoints where student thinking is surfaced and examined.

Counselors can open direct conversations with students about how they are starting their work and what they do when they feel stuck. Many students are relying on AI at that exact moment. Naming that decision point helps them see it more clearly.

Students are not losing their ability to think. They are responding to the environment around them. When schools make space for independent thinking alongside AI use, students adjust quickly. The capacity is still there. It needs to be practiced.

This is one of many hidden shifts AI is introducing into education. I write these newsletters to help educators and school leaders see them early and respond intentionally. If that matters to you, stay with the series.