AI, Education, and Cheating: Problem or Educational Opportunity? Banning isn’t the answer.

Table of Contents

1. Artificial Intelligence unmasks a silent crisis
2. Why banning ChatGPT is not enough: the crisis of instructional design
3. The thin line: academic integrity in the age of AI
4. The solution is not surveillance, but redesign: anti-cheating factors
5. Concrete proposals: a simple framework for teachers and schools
6. AI is the future. Educating is our responsibility

Artificial Intelligence unmasks a silent crisis

The introduction of Generative Artificial Intelligence (AI) tools like ChatGPT into classrooms has sparked a heated debate. Concerns about school cheating (a set of unfair and opportunistic behaviors) are legitimate and widespread, leading many institutions to consider an absolute ban.

However, the central thesis emerging from scientific research is clear: AI does not “create” the problem of cheating; it simply makes it more visible, exposing the structural flaws of the traditional evaluation system.

The true opportunity is not to fight the wave, but to learn how to ride it. The solution lies in the radical redesign of educational paths and assessments, not in the demonization of a tool.

Why banning ChatGPT is not enough: the crisis of instructional design

The temptation to ban tools like ChatGPT or Copilot is understandable. If an essay can be written in a few seconds, why should the student struggle for hours?

This question reveals an uncomfortable truth: cheating already existed. The pressure for grades, the perception of irrelevant assignments, or simply a lack of time have always pushed some students to copy, commission, or cheat.

AI is not the cause, but the catalyst that exposes the fragility of tasks based on memorization or trivial synthesis.

An analysis of “Real student behaviors” (as suggested by recent studies in Frontiers in Education) shows that students use AI for ambivalent reasons. The research distinguishes two main uses:

• Legitimate Use: Brainstorming, obtaining complex explanations, or correcting grammar.
• Illicit Use: Complete generation of assignments.

The greatest push toward cheating stems from a combination of performance pressure, ethical ambiguity and, above all, a lack of clear and shared rules.

If a task can be completed entirely by a machine, perhaps the problem is not the machine, but the educational value of the task itself.

The thin line: academic integrity in the age of AI

Defining what is permissible and what constitutes cheating in the age of AI is a conceptual challenge that requires a rapid ethical and pedagogical update.

A systematic review on “academic integrity in the age of AI” (published in ScienceDirect) highlights that the boundary between legitimate and illicit use is often fluid and unclear.

To facilitate understanding, we can distinguish the following cases:

• Legitimate Use (Support): this includes actions such as asking the AI to explain a complex concept, brainstorming ideas or titles, or generating data for documentation (if cited).
• Gray Zone (Excessive Assistance): this includes using AI to rewrite passages to improve clarity or to create a first structural draft.
• Illicit Use (Cheating): it is considered cheating to have the AI generate the entire paper without citation or integration, or to submit the raw AI output as one’s own.

This conceptual framework pushes the educational community to call for an urgent rethinking of:

1. Evaluation rubrics: we must ask ourselves whether we are evaluating the ability to produce a text or the ability to direct AI to produce a quality text.
2. Ethical training: it is necessary to integrate the concept of the ethical prompter and the documentation of AI use.

The solution is not surveillance, but redesign: anti-cheating factors

Research is unanimous: the winning approach to reducing the temptation to cheat is one that aims to raise the intrinsic value of learning, making cheating less attractive and more difficult.

Recent studies on assessments (such as those highlighted by eSchool News) show that cheating decreases drastically with the implementation of three fundamental pillars:

A. Continuous formative assessment (Assessment for learning)

When the focus shifts from the single final grade to the learning process, students have less incentive to cheat. Formative assessment (or assessment for learning) is continuous, aims to provide feedback for improvement, and does not penalize error but uses it as a teaching tool. This reduces pressure and performance anxiety.

B. Authentic tasks and real-world problem-solving

Cheating disappears when tasks require critical application, a unique personal synthesis, or the resolution of complex problems that do not have a single answer found online. Let’s look at an example:

• Developing a sustainability plan for one’s own school (requiring local data).
• Critical analysis and comparison between AI output and an original text, justifying the differences.
• Simulations and role-playing that cannot be replicated by a generic prompt.

C. Greater agency and co-creation

Giving students greater agency means involving them in the choice of the topic, the assessment format, and even the rules for using AI. When students feel that the task is theirs and reflects their interests, engagement and integrity increase.

Concrete proposals: a simple framework for teachers and schools

Approaching AI with a clear strategy is the key to transforming the challenge into an opportunity. Schools and teachers should move toward the development of a simple and transparent framework:

Define the rules of engagement (A.I. Use Policy)

It is essential to establish a clear “AI Use Statement,” dividing use into defined categories:

• What is allowed (enhancement): clarify that AI can be used as personalized tutoring, for brainstorming, for linguistic proofreading, or for simulating scenarios.
• What is forbidden (plagiarism/cheating): define as cheating the submission of output entirely or largely generated by AI as original work without documentation.
• How to document (transparency): ask students to include a note in their work, specifying which tools they used and for which phase of the task (e.g. “Used ChatGPT to generate 5 possible titles; the text is entirely my own”).

Concrete examples of positive AI use in class

AI, if used correctly, is an unprecedented learning tool:

• Personalized tutoring: AI can explain complex concepts in different ways until the student grasps them (like a 24/7 tutor).
• Linguistic feedback and revision: it helps students improve writing quality in real time, far beyond a simple spell checker.

Complex simulations: AI can generate scenarios, data, and personas for history, economics, or science simulations.

AI is the future. Educating is our responsibility

Generative Artificial Intelligence is an epochal change, not a simple tool to be banned. Research shows us the way: education must evolve with technology.

Stopping the evaluation of only the final product (the essay) and starting to value the process (the ability to manage and direct AI) is the only sustainable strategy.

It is not AI that puts academic integrity at risk, but our reluctance to redesign an evaluation system that was already obsolete. The opportunity is to train a new generation of critical thinkers who know not only how to use technology, but also how to master it with ethics and transparency.