Direct conversations – New findings show only about 1 in 3 students say their school has a school-wide AI policy, while most students report policies differ by teacher. The recommendations center on direct, school-wide conversations—with students and staff—to set guardrails, protect
For years, teachers have been trained to pull students into “meaningful conversations” about the real world—asking thoughtful questions, listening closely, and taking students seriously. But in many classrooms, that lesson still stops short when the subject is AI.
New findings released by RAND’s American Youth Panel suggest why. Only about 1 in 3 students say their school has a school-wide policy on the use of AI. Many students also say AI policy varies by teacher. In the same findings. 67 percent of students endorsed the statement: “The more students use AI for their schoolwork. the more it will harm their critical thinking skills.” The RAND report’s recommendation is bluntly practical: schools should hold direct conversations with students about AI.
The numbers connect to a wider reality inside schools. The Center for Democracy and Technology reports that approximately 85 percent of teachers and students say they use AI for schoolwork. That means many students aren’t waiting for a formal policy before encountering AI—they’re navigating it anyway. often without consistent guidance from adults.
That inconsistency is where the stakes land. If policies change depending on the teacher. students can end up learning one thing in one class and something else in another. And if most students already worry that frequent AI use could weaken critical thinking. then the missing piece isn’t just a document. It’s a shared conversation about what schools believe AI should—and shouldn’t—do.
The proposed starting point begins with staff meetings that don’t treat AI as an outside add-on. Teachers and school leaders are urged to talk through questions such as: Is the goal to make things easier for students or for teachers—and is that what the school wants?. If making things easier is part of the plan, when does that help, and when does it cross a line?.
The discussion is also expected to move beyond convenience. How can schools implement AI and LLM tools so the learning community benefits—through increased efficiency. time savings. and the ability to gather and analyze more data?. What guardrails can protect learning experiences that many schools value. including productive struggle. working through complex problems. and devising. testing. and refining solutions?.
Students are central to that conversation, not just as recipients of rules. Educators are encouraged to ask how students will be taught to critically analyze the information and “answers” AI tools provide. They’re also pressed to consider how skillful students are at identifying bias—whether students will ask what the source is for information. what perspective a source has. and whether they can distinguish fact from opinion. The example given compares the distance between Earth and the sun (fact) with the filibuster as a tool for promoting democracy (opinion).
Just as important, the questions look inward at who drives innovation in schools. What skills do students—and teachers—need to strengthen so they remain the drivers of AI innovation?. And where can schools look for models?. The guidance includes asking whether other schools or people they trust and respect have already implemented AI policies. and what could be learned from them.
Even the process of deciding matters. Schools are prompted to ask what systems they have—or can build—to include student voice in determining when and how AI gets used.
With students. the recommended conversation is framed around values rather than only rules: what is valuable about the work students do together at school. and how might AI increase that value or undermine it?. There is also a focus on integrity—what it means “to us. as individuals and as a school. ” and how AI use can be designed to support it.
The questions students are asked to consider are simple enough to start immediately: What do they know about AI? What do they want to know about it? What are possible ways AI might be used in schoolwork, and what benefits and drawbacks do they see?
This is also positioned as more than a one-time meeting. An AI policy, the guidance warns, cannot just be overlaid while the school continues business as usual. AI is described as a powerful tool with the ability to disrupt—disruption that can be beneficial. including disrupting inequitable access to information and learning tools. but also harmful if it fuels complacency and undermines critical thinking and curiosity.
That’s why the policy must be aligned with school values. backed by “thoughtful. school-wide conversations” built around those shared beliefs. During those conversations. staff are encouraged to use the phrase “I don’t know. ” as a way of modeling uncertainty tolerance—because educators do not yet have all the answers about what AI can or should do. and how it might support or undermine critical thinking and curiosity.
The end goal isn’t to eliminate uncertainty. It’s to teach students how to puzzle through complex issues. In the guidance. teachers are told that by engaging students in the questions laid out—about guardrails. bias. integrity. and critical analysis—they’re modeling how to handle complex problems at the highest level.
In a school climate where only about 1 in 3 students report having a school-wide AI policy. and where 85 percent of teachers and students say they use AI for schoolwork. the most immediate change may not be an updated rulebook. It may be the decision to talk—directly. school-wide. and in the open—about what AI is doing to learning. and what the community wants learning to remain.
AI policy in schools RAND American Youth Panel student critical thinking AI in education LLM tools school values student voice teacher conversations Center for Democracy and Technology