Talk About AI Without Losing Your Students’ Trust

The AI Conversation Is Already Happening

Your students are already talking about artificial intelligence.

They are comparing tools, testing prompts, sharing shortcuts, and trying to understand what is permitted. Some are genuinely curious about how AI can improve their learning. Others are worried that using it in any form could place them under suspicion.

The question is not whether students will encounter AI. The question is whether their school will help them develop the judgement required to use it responsibly.

This is where many classroom conversations go wrong.

Teachers often begin with warnings. They explain what students must not do, how misuse might be detected, and what consequences could follow. The intention is understandable. Schools must protect academic integrity and ensure that submitted work represents a student’s genuine thinking.

However, when the first conversation sounds like an investigation, students learn to hide their AI use rather than discuss it honestly.

A more effective approach begins with trust.

Students need to understand that the purpose of an AI policy is not to catch them using technology. It is to help them recognise when technology supports learning and when it replaces the thinking they are expected to demonstrate.

Start With Learning, Not Misconduct

The first classroom conversation about AI should not begin with cheating.

It should begin with learning.

Ask students what they believe AI can do well. Then ask what it does poorly. Encourage them to discuss situations in which an AI tool might help them understand a difficult concept, generate questions, improve organisation, or explore another perspective.

After that, ask where the risks appear.

Students may identify inaccurate information, invented sources, cultural bias, shallow explanations, privacy concerns, or the temptation to submit work they did not create.

This sequence matters.

When students first examine both the possibilities and the limitations of AI, the conversation becomes intellectually honest. It no longer feels like a lecture designed to restrict them.

The teacher can then connect AI use to familiar IB expectations such as inquiry, reflection, communication, critical thinking, and academic integrity.

The message becomes clear.

AI is not automatically good or bad. Its value depends on how, why, and when it is used.

Explain the Difference Between Support and Substitution

Students often struggle with AI policies because the language is too broad.

They may hear that AI is allowed for brainstorming but not for writing. Yet the boundary between brainstorming and writing can feel unclear. They may use a tool to generate an outline, revise a sentence, suggest examples, or explain a concept without knowing whether they have crossed a line.

Teachers need to make the distinction concrete.

AI supports learning when it helps the student think.

AI substitutes learning when it does the thinking the student is expected to demonstrate.

For example, asking AI to explain the difference between correlation and causation can support understanding. Asking it to write the student’s analysis of a data set replaces the analysis being assessed.

Using AI to generate practice questions can support revision. Using it to produce answers for a graded task substitutes the student’s knowledge.

Requesting feedback on whether an argument is clear may support reflection. Asking AI to rewrite an entire essay until the student’s original voice disappears may replace authorship.

Students understand policies better when they can compare realistic examples instead of memorising abstract rules.

Use Three Questions for Every AI Decision

A simple decision framework can help students judge whether an AI use is appropriate.

Before using a tool, students should ask three questions.

What am I expected to learn?

The student must first identify the purpose of the task.

If the task is designed to assess research, interpretation, writing, problem solving, or reflection, the student must complete the thinking connected to that skill.

A tool should not remove the exact challenge the task was designed to measure.

What work is the AI doing?

Students should describe the role of the tool honestly.

Is it explaining a concept, suggesting questions, summarising notes, generating wording, producing an argument, analysing evidence, or creating a final response?

The more intellectual work the tool performs, the greater the risk that it is replacing the student’s own contribution.

Could I explain my process openly?

This is one of the strongest tests of responsible use.

Students should be comfortable showing their teacher what they entered, what the tool produced, what they accepted, what they rejected, and how the final work changed.

When students feel the need to hide their process, they probably already recognise that the use may be inappropriate.

These questions make the policy practical. They move the conversation away from fear and towards judgement.

Avoid Turning Teachers Into Detectives

A surveillance based classroom damages trust quickly.

When teachers repeatedly warn students that they can identify AI writing, students may feel that every polished sentence will be treated as evidence against them. Strong writers can become anxious. Multilingual learners may worry that language support tools will make their work appear suspicious.

Teachers should not ignore genuine concerns about authenticity. However, the response should focus on understanding the student’s process rather than making accusations based on style.

A productive conversation might begin with questions such as:

“Can you walk me through how you developed this argument?”

“Which sources shaped this paragraph?”

“How did your first draft change?”

“What part of this response was most difficult to write?”

“Did you use any digital tools while preparing this work?”

These questions invite explanation. They also reveal whether the student understands the content and can account for the decisions behind the final submission.

Process based assessment is usually more educational than attempting to identify misconduct from a finished document alone.

Make Disclosure Normal

Students are more likely to disclose AI use when disclosure is treated as a normal part of academic practice.

A school can create a simple AI use statement that students attach to relevant assignments. The statement should not imply wrongdoing. It should document the learning process.

Students might explain which tool they used, why they used it, the prompts they entered, the output they received, and how they evaluated or changed that output.

A student who used no AI should also be able to state that clearly.

The goal is transparency, not confession.

Teachers can model this behaviour by sharing their own responsible use of AI. For example, a teacher might explain that they used a tool to generate sample misconceptions, create practice questions, or compare possible lesson structures.

They should also explain what they checked, corrected, and rejected.

When teachers reveal their own evaluation process, students see that responsible AI use requires human judgement. They also understand that transparency applies to the whole learning community, not only to students.

Separate Guidance From Punishment

An AI policy should help students make decisions before it explains what happens after misconduct.

When rules about AI appear only inside an academic integrity document, students may associate every use of the technology with suspicion and punishment.

Schools can reduce this effect by creating separate classroom guidance that explains appropriate use in plain language.

The academic integrity policy can still define unacceptable conduct and formal consequences. However, the learning focused AI guidance should answer the questions students face during ordinary schoolwork.

It should explain:

1. Which uses are generally encouraged

2. Which uses require teacher permission

3. Which uses must be acknowledged

4. Which uses are not permitted

5. How expectations change across subjects and assessments

6. What students should do when they are uncertain

This separation helps students see that the school is not trying to ban exploration. It is trying to establish responsible boundaries.

Be Specific About Different Tasks

A single rule cannot cover every IB subject or assessment.

AI use that is acceptable during early brainstorming may be inappropriate during a final assessed task. A tool that supports language development may interfere with an assessment designed to measure independent writing. A teacher may permit AI generated practice data but prohibit AI analysis of the student’s actual investigation.

Teachers should explain expectations at the beginning of each task.

Students need to know whether AI can be used for topic exploration, planning, research support, vocabulary development, feedback, editing, coding, image generation, translation, or reflection.

They also need to know which parts of the task must remain completely independent.

A simple task specific statement can prevent confusion:

“You may use AI to generate practice questions and clarify concepts. You may not use it to write, rewrite, or analyse any part of your final response. Any AI use must be recorded in your process log.”

Clear instructions are more effective than expecting students to interpret a general school policy every time they begin an assignment.

Acknowledge That Different Students Need Different Support

AI conversations must also consider equity.

Some students have access to paid tools, stronger devices, private tutoring, and reliable internet. Others may depend entirely on free tools with limited capabilities.

Students with additional learning needs may use AI supported features for organisation, reading, translation, speech recognition, or language development. These uses should not automatically be treated in the same way as generating assessed content.

Schools need to distinguish between accessibility support and intellectual substitution.

This requires careful coordination among teachers, learning support teams, coordinators, students, and families. The aim should be to protect fair assessment while ensuring that students are not denied legitimate support.

An equitable policy does not simply give every student the same rule. It ensures that all students can demonstrate their abilities under fair and clearly explained conditions.

Teach Students to Challenge AI Outputs

Responsible AI education cannot stop at permission and citation.

Students must learn how to evaluate what a tool produces.

An AI system can generate confident statements that are inaccurate, incomplete, biased, or entirely invented. It may simplify complex debates, ignore local contexts, reproduce stereotypes, or provide sources that do not exist.

Teachers can turn these weaknesses into learning opportunities.

Give students an AI generated response and ask them to identify assumptions, missing perspectives, unsupported claims, and factual errors. Ask them to improve the response using reliable evidence.

Students can also compare answers produced by different prompts. They can examine how wording influences the result and discuss why a fluent response is not necessarily a trustworthy one.

This approach develops critical thinking rather than passive dependence.

The student’s role is not to accept AI output. It is to question, test, verify, and decide.

Use Language That Preserves Trust

The words teachers choose shape how students respond.

Statements such as “We know when you have used AI” create defensiveness. Statements such as “You will be punished if your work sounds artificial” create anxiety without providing useful guidance.

Trust building language is different.

“We want you to learn how to use these tools responsibly.”

“Your thinking must remain visible in your work.”

“Ask before using AI when the task instructions are unclear.”

“You will not be penalised for discussing a mistake honestly.”

“Transparency helps us support you.”

This language does not weaken academic standards. It makes those standards easier to understand and follow.

Students are more likely to ask for help when they believe the teacher will respond with guidance rather than immediate suspicion.

Respond to Misuse as a Learning Moment

Not every inappropriate use of AI begins with an intention to cheat.

A student may misunderstand the rules, rely too heavily on a rewriting tool, or assume that generated text is acceptable because they changed a few words.

Schools still need consistent consequences for serious or deliberate misconduct. However, teachers should distinguish between confusion, poor judgement, and intentional deception.

When appropriate, the response can include a reflective conversation.

What did the student ask the tool to do?

Why did they use it?

Which part of the policy did they misunderstand?

What learning was lost because of the tool?

How could they complete the task differently?

This does not mean removing accountability. It means ensuring that accountability produces better future decisions.

A student who only receives a penalty may learn to hide AI use more carefully. A student who understands the academic problem is more likely to change their behaviour.

Include Students in Policy Development

Students can provide valuable insight into how AI is actually being used.

Schools should invite them to discuss the tools they encounter, the pressures that lead to misuse, and the areas where current rules feel unclear.

Student involvement does not mean allowing learners to decide whether academic standards apply. It means using their experience to make guidance realistic and understandable.

A student advisory group can review sample scenarios, suggest clearer language, and identify emerging tools before they create widespread confusion.

This process also builds ownership.

Students are more likely to respect a policy when they understand its purpose and have contributed to how it is communicated.

A Practical Classroom Conversation

Teachers do not need a dramatic assembly or a threatening presentation to begin the discussion.

A productive classroom conversation can follow a simple sequence.

First, ask students how they currently see AI being used in education.

Second, discuss useful applications and genuine risks.

Third, explain the difference between support and substitution.

Fourth, work through subject specific examples.

Fifth, introduce the three decision questions.

Sixth, explain how students should disclose their use.

Seventh, invite questions without attaching suspicion to them.

End by making one promise clear:

Students who ask about an AI use before submitting work will receive guidance, not punishment for asking.

That promise creates the psychological safety required for honest conversations.

Conclusion

Students do not need another vague warning about artificial intelligence.

They need practical guidance, realistic examples, and teachers who are willing to discuss the technology without treating every question as evidence of misconduct.

The strongest AI conversations protect both trust and academic integrity. They make clear that students remain responsible for their thinking, their evidence, their decisions, and the work they submit.

At the same time, they recognise that AI will remain part of education and professional life.

The goal is not to create students who are afraid to use technology. It is to develop students who know when a tool supports learning, when it replaces learning, and how to make that distinction honestly.

When schools lead with clarity instead of suspicion, students are more likely to disclose their process, ask better questions, and take ownership of their work.

That is not weaker academic integrity.

It is academic integrity taught well.