We Are Teaching AI Literacy to Students. Who Is Teaching the Adults?

Schools are beginning to expect students to understand AI, question what it produces, protect their information, recognize bias, explain when they used it, and know when the technology should not do the thinking for them. Those are reasonable expectations. Students need more than access to AI. They need the judgment to use it without handing over decisions, learning, or responsibility they still need to carry themselves.

The harder question is whether the adults asking students to do all of that have been given enough opportunity to develop the same understanding.

A teacher may be expected to explain why an AI answer should not automatically be trusted while still feeling unsure about how the system produced that answer. A principal may be responsible for setting classroom expectations without knowing what information teachers are entering into different tools. An instructional coach may be asked to help teachers redesign assignments around AI while still working through when AI support strengthens student thinking and when it quietly replaces it. A district leader may be sitting through a vendor presentation and trying to decide whether a product is appropriate for students without having a shared set of questions for evaluating what the system does, what information it receives, or where human judgment still belongs.

None of those situations means the adults involved are unprepared professionals. It means the work changed before the professional learning caught up.

Schools are asking educators to teach, supervise, evaluate, and govern AI use while many teachers and leaders are still building the knowledge and judgment those responsibilities require.

That gap is easy to underestimate because adult AI literacy can look deceptively simple from the outside. A school may provide a professional development session where educators learn how to generate a lesson plan, draft an email, adjust a reading passage, or create a rubric. Teachers leave knowing several things the technology can do, and the district can reasonably say that AI professional learning has begun.

But knowing how to use an AI tool is only one part of being literate enough to make professional decisions with it.

A teacher still needs to know why a confident response can be wrong, what information should not be placed into a system, how bias can appear in an output, why the same request may produce different responses, and when an AI-generated scaffold removes too much of the thinking a student is supposed to do. The teacher needs to know when AI is useful for creating a starting point and when the task depends too heavily on knowledge of a particular student for a general system to make the decision.

The teacher also needs enough understanding to explain those distinctions to students.

Consider what this can look like in a high school classroom. A teacher tells students that AI may be used for brainstorming but not for writing the final response. A student asks whether using AI to reorganize an argument counts as brainstorming. Another asks whether receiving feedback on a draft is allowed. A third asks whether using AI to suggest stronger evidence is different from asking it to write the paragraph.

Those are not questions a simple allowed-or-prohibited rule can answer. The teacher needs to understand what the assignment is intended to develop and which uses of AI preserve that learning. If the goal is argument construction, allowing the system to generate the argument may remove the very thinking students need to practice. If students have already developed their claims and are evaluating AI-generated counterarguments, the same tool may deepen the work rather than replace it.

The teacher does not need advanced technical knowledge to make that decision. The teacher does need AI literacy.

The same issue appears when teachers are expected to help students evaluate AI-generated information. A student brings an AI response into class and says the tool provided three sources supporting a claim. The teacher may recognize that the information needs to be verified, but students need more than an instruction to "check the AI." They need adults who can show them what checking actually looks like.

Which claims need another source? How does a student determine whether a cited source exists? What happens when an answer sounds reasonable but the evidence does not support it? How should students respond when a tool gives one answer today and a different answer another time?

Students learn those habits partly from what educators model. If adults treat AI output as probably correct unless an obvious error appears, students can learn that habit too. If adults routinely question claims, check sources, compare information, and explain why they reject a response, students see what responsible use actually looks like.

That is one reason AI literacy cannot be assigned only to students. The adults establish the culture around it.

The issue becomes even more important for school leaders because leaders are making decisions that affect far more than one classroom. A principal may need to determine whether teachers can use AI to draft family communication, whether AI-supported feedback belongs in instruction, how staff should respond when a student uses AI in an unexpected way, or what professional expectations should apply when teachers choose not to use AI at all.

A district leader may have to decide whether a new product should be purchased, what evidence a vendor should provide, what student information the system can receive, which uses require human review, and whether a tool that appears efficient actually improves anything the district cares about. Those are AI literacy decisions too, even though they may look like purchasing, policy, privacy, or instructional decisions on the surface.

A leader who does not understand the technology well enough to question its claims can unintentionally turn vendor language into district practice. A leader who understands privacy but not instruction may approve a product that protects data while weakening the learning the school hoped to improve. Another leader may understand the classroom value but miss the implications of entering identifiable student information into a system that was never reviewed for that use.

No single person has to know everything. The school does need enough shared literacy across roles that important questions do not disappear between departments.

This is also where some professional development falls short. AI training can become a demonstration of what the newest tools can produce. Educators learn how to generate a lesson, summarize a document, create a presentation, or save time on routine work. Those skills may be useful, particularly for teachers already carrying more work than the school day can hold.

But an educator can become very skilled at using AI without becoming more skilled at deciding when it belongs in the work.

That distinction matters because AI literacy also includes the ability to decide not to use AI.

A teacher who understands what a tool can do and still chooses to write a sensitive family communication personally may be demonstrating strong judgment. A counselor who knows how AI can organize information but decides that a particular student situation contains too much personal context for the tool may be demonstrating AI literacy. A special educator who recognizes that an AI-generated recommendation does not account for what is known about a particular learner and rejects it is not failing to use the technology effectively.

The decision not to use the tool can be evidence that the adult understands it.

That is a very different standard from counting how often educators use AI.

Why It Matters

Schools cannot build student AI literacy on top of adult uncertainty indefinitely. Students will ask questions that policies do not anticipate. They will find uses teachers have not seen before. Tools will change in the middle of the year. New features will appear inside products educators have used for years. Families will ask why one teacher permits a particular use while another does not. Teachers will encounter student work that looks different from anything they have graded before.

In those moments, the school cannot rely only on a list of approved tools. It has to rely on adult judgment.

That judgment becomes more dependable when educators share enough knowledge to recognize what they are actually deciding. An AI-literate teacher should understand enough about generative AI to know why an output can sound convincing and still be wrong. The teacher should know that entering information into a prompt is also a decision about data. The teacher should be able to examine whether AI is supporting a learning goal or performing part of the skill students are supposed to develop. The teacher should also be able to explain why one use is appropriate in a particular assignment and another is not.

That does not mean every teacher needs to become an AI specialist.

A third-grade teacher does not need to understand the technology at the same technical level as someone who designs AI systems. An English teacher does not need the same depth of technical knowledge as a computer science teacher. A counselor does not need to become an instructional technology specialist.

They do need enough literacy for the decisions their roles require.

School leaders need another layer. They need to understand enough to evaluate what teachers are being asked to do, what risks the school is accepting, what questions should be asked before a tool is approved, and whether policies reflect what is actually happening in classrooms. They also need enough literacy to recognize when the answer should come from someone else.

A principal should know when a question about student information needs privacy or technology expertise. A curriculum leader should know when an instructional use has implications for students with disabilities. A technology department should know when a technically secure product raises questions about learning that cannot be answered through a security review.

AI literacy at the system level includes knowing where expertise lives and when a decision needs more than one perspective.

This is why one professional development session will not solve the problem. Adults build literacy by using knowledge repeatedly in real decisions.

A teacher becomes more literate when examining an AI-generated resource and noticing that the reading level changed but the concept became weaker. A principal becomes more literate when reviewing a proposed classroom use and asking what student information enters the tool and what the teacher still reviews. An instructional coach becomes more literate when helping a teacher decide whether AI feedback preserves the thinking students need to do.

Those experiences need shared language around them.

Schools can begin building that language without creating another large initiative. At a faculty meeting this month, leaders can move away from a tool demonstration and give staff one realistic school situation to examine together. A teacher wants to use AI to create differentiated reading materials. What does the teacher need to check before those materials reach students? What information can safely be entered? How would the team know whether the adaptation improved access without removing the intended learning?

The purpose is not to arrive at one perfect answer. The purpose is to practice the judgment educators will need when the next tool appears.

Grade-level and department teams can bring the same work into planning. Once during the month, teachers can choose one AI-supported assignment or resource and discuss where the tool contributed, where the educator's judgment was still required, and what would have made the same use inappropriate.

That begins turning AI literacy from something educators know into something they can apply.

Instructional coaches can make the work more specific by helping teachers examine AI use through the learning goal. Instead of beginning with what the tool can do, the conversation can begin with what students are supposed to understand or demonstrate. From there, the teacher and coach can decide whether AI removes a barrier, strengthens the learning, or takes over part of the thinking the lesson was designed to develop.

Principals and district leaders can do their own version during leadership meetings. They can take one real AI decision already facing the school and examine it from several sides. What is the instructional value? What information enters the system? What could go wrong? What needs human review? Which students might experience the tool differently? What evidence would show that the use is helping?

Leaders cannot expect teachers to develop thoughtful AI judgment while leadership conversations stay at the level of tools and adoption. The adults making those decisions need practice too.

Districts can also make professional learning more useful by establishing a small baseline of AI literacy that every educator should have while allowing different roles to develop additional knowledge based on their responsibilities. The baseline does not need to become another certification or complicated framework. Educators should be able to explain in plain language what AI can and cannot reliably do, recognize when student information or context makes a use more sensitive, question an output rather than automatically trust it, understand the school's expectations, and know where to go when the decision exceeds their role.

From there, the learning becomes more specific.

Teachers need to understand AI in relation to instruction and assessment. Counselors need to understand it in relation to privacy, student voice, and sensitive information. Special educators need to understand how AI-generated supports interact with individual needs and required services. Principals need to understand enough across those areas to lead consistent practice. District leaders need to understand how instructional value, privacy, cybersecurity, purchasing, equity, and professional judgment connect.

That is what adult AI literacy begins to look like inside a school system.

It is not everyone knowing the same amount. It is everyone knowing enough to carry the responsibility attached to the role.

Schools should also be careful about creating an expectation that adults become fully confident before they talk with students about AI. That standard is unrealistic because the technology will continue changing. Teachers will encounter questions they cannot immediately answer, and leaders will sometimes have to make decisions with incomplete information.

AI literacy does not mean certainty. It means knowing how to respond when certainty is not available.

An AI-literate educator can say that an answer needs verification. A leader can pause a decision because a privacy question has not been answered. A teacher can tell students that the class needs to investigate why two AI outputs disagree. A principal can acknowledge that a policy does not yet address a new use and bring the question back to the people who should help decide it.

Those are not signs that the adults are behind. They are signs that the adults know how to think responsibly in a changing environment.

That may be the most important part of the work.

Students do not need adults who can predict every new AI tool they will encounter. They need adults who know what questions to ask when the tool changes. They need teachers who can tell the difference between assistance and replaced thinking. They need leaders who can explain why a boundary exists rather than simply announce one. They need counselors and student-support staff who recognize when sensitive human context should not be handed to a system. They need adults who can question a polished answer, protect information, notice bias, and remain responsible for decisions made with technological assistance.

Schools are right to teach AI literacy to students, but that responsibility cannot begin and end with students. The adults shaping assignments, writing policies, purchasing tools, answering families, protecting student information, and deciding what responsible use looks like need their own continuing opportunities to learn.

The strongest AI literacy culture in a school may not be the one where everyone uses AI most often. It may be the one where adults can explain when to use it, when to question it, when to stop, and why.

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.