Student AI Misuse: The 3 Solutions People Overlook

Here's a situation most of your students will find themselves in at some stage this term. They get handed an assignment in class, come home, and sit down to start it. It's a big task, and it's going to take considerable time to do properly. They look at the question and realise that most of that work would disappear if they simply put it into AI instead. A big task suddenly becomes a much smaller one. Standing at that ethical fork in the road, they pause: take the honest route and do the work themselves, or take the easy route and let AI do it for them?

In this moment of ethical dilemma, one of the key forces that determines the choice our hypothetical student makes is the school's culture. Your school's culture serves to either act as a constraining force, bringing students back from the ledge of academic misconduct, or as the enabler that sends them over it.

The difficulty with conversations about culture is that it's a vague term which means different things to different people, and defining the practical, real characteristics of a school's culture, especially what it actually is in reality as opposed to what we want it to be, is even harder still. We define culture as the beliefs, values and actions of a school that function as its unwritten rules. If you want to understand why teachers, students or parents act the way they do, one of the first places to look is those unwritten rules.

We've found there are 3 critical components of a school's culture that constrain or encourage students to use AI in the wrong ways.

Lever 1: Student Perception of Peer Behaviour

Members of any organisation are constantly assessing two types of norms to determine how they should act. First, they assess injunctive norms: what they believe people should do. Traditionally, injunctive norms are defined by the rules. But they also assess descriptive norms: what other people are actually doing. When there's a significant gap between the two, people are more likely to follow the crowd and do the wrong thing rather than follow the rules. Someone driving on a highway knows the speed limit, the injunctive norm, but if they observe everyone else speeding, they'll speed up too.

The same is true of students. They may know the school's rules around AI and understand their teacher's guidance, but they'll also look at their peers and how closely those peers are actually following the rules. Where the gap is significant, students are encouraged to follow the crowd. Doing the right thing in this situation can even feel unfair: a student thinks, “What's the point of following the rules if everyone else is cheating and I'm the only one constrained by them?” Misusing AI doesn't just become easy to rationalise. It starts to feel like the fair thing to do.

For schools trying to understand and address misuse, it becomes critical to understand how students perceive their own habits, that is, whether they see themselves as doing the right thing, and how they perceive their peers' habits. This is the ultimate barometer for predicting AI misuse, and it points directly to the first set of remedial actions schools should take.

One further point worth making here: without clear injunctive norms in the first place, the only way for someone to assess what to do is by looking at what other people are doing. Without speed limit signs, drivers assess speed purely by reference to other drivers. Without clear rules on AI use, students will do the same. This is why a clear AI use policy matters so much. We recently checked the AI policies of 20 of our client schools and found that not one had a publicly facing policy. Their AI policy either lived scattered across other policies, like technology or student behaviour policies, or existed but wasn't public. A school's first step, before worrying about descriptive norms at all, is making sure it has a concise, publicly available AI policy, and that this policy is communicated regularly, ideally as clear guidance every time a task is assigned.

Lever 2: Consistency Across Teachers

Members of an organisation also judge how seriously to take the rules by watching how people in power act. Do those in power reference, follow, and uphold the rules? If not, members are encouraged to push the limits themselves.

The same is true of students. They constantly assess how teachers discuss AI, uphold expectations, and set assignments. Where teacher practices are consistent with school policy, and consistent with each other, students understand the rules are real, will be applied, and can't be gamed. Inconsistent expectations communicate the opposite: that the rules are lip service. This lets students adjust their behaviour accordingly. “Teacher A applies the rules, so I need to do the right thing. Teacher B doesn't, so I can cut corners there.”

Practically, this means there are three metrics schools should track around teacher consistency.

• Teachers' buy-in to the rules. Do teachers actually agree with the school's position on AI? If not, they're unlikely to enforce it rigorously, and misuse cases are unlikely to fall. Building buy-in starts with a clear mission for why AI matters to students' futures, communicated regularly, alongside giving staff a genuine voice in shaping the rules.

• Teachers' perceptions of what other teachers are doing. As with students, if teachers feel that no one else is really applying the rules or expectations, they have less incentive to do so themselves.

• Students' perceptions of what their teachers are doing. The first two are lead indicators. This is the lag indicator. If students believe teachers apply the rules inconsistently, that's the signal to investigate the first two and address what's actually happening.

  
  

Schools can assess these indicators by surveying students and teachers at one or two points across the year. Use student perceptions to flag whether a consistency issue exists in the first place, then use the two teacher indicators to work out how to actually solve it.

Lever 3: Recognition and Reward

Teachers have two tools to motivate students: the stick, rules and punishment, and the carrot, recognition and incentive. The stick is the easier of the two to use. It takes little creativity, and it's usually the first tool people reach for when they're new to a problem or uncertain how to approach it. That's probably a fair description of where most schools sit today when it comes to AI: rules or outright bans, and punishment when those rules are broken.

But most people would also agree that while the stick is the easier motivator to reach for, it's the carrot that provides the more sustainable and longer lasting motivation, even though it's the tool schools tend to use least in the early days of navigating AI. In this context, the carrot means recognising and rewarding good AI use. Ultimately, the biggest lever for getting students to use AI effectively and ethically is making sure that use is actually seen and rewarded by teachers.

The challenge for schools is that using the carrot takes more creativity and thought than simply falling back on rules and punishment. The best examples we've seen of effective recognition and reward are built directly into a school's assessment redesign process. As we've covered in an earlier newsletter, great assessment redesign isn't about AI-proofing a task. It's about redesigning it so students learn and apply subject content while also developing genuine AI skills. Most schools who get this right have built AI disclosure processes. These are more than a box-ticking exercise or an AI-flavoured bibliography.

Instead, the disclosure form becomes a second work product, marked alongside the assignment itself, ideally carrying similar weight. In it, a student explains how they applied their AI skills, how they brought critical evaluation to the outputs AI generated for them, and how they used AI creatively to approach the task. The form is then marked against an AI literacy rubric, so a student receives their normal project mark, alongside a separate AI literacy mark.

Disclosure lets good AI use be recognised in the first place. From there, it's just a matter of applying the standard tools to reward it.

How Elevate can help. We've spent the last 18 months surveying schools and students around the world to understand the challenges each is facing with AI and how best to respond, and we've used that data to build a diagnostic tool for unpacking the black box of school AI culture, covering exactly the three levers above. Want to see exactly where your school sits? Try a free module of our AI Culture Snapshot, benchmarked against the schools we work with, covering how your students perceive their peers, how consistent your teachers actually are, and how well good AI use gets recognised: Try a free module