Psychological Determinants of Academic Integrity in the Use of Generative AI in Higher Education

In the modern classroom, a quiet revolution has taken hold, not with the clatter of new machinery, but with the silent hum of a computer screen. For decades, educators have relied on the idea that a student's submitted work is a direct reflection of their own mind, a unique product of their effort and understanding. This concept of authorship is the bedrock of academic trust. However, a new class of tools known as generative artificial intelligence has begun to blur the line between a helpful assistant and a ghostwriter. These systems can summarize complex texts, draft essays, write computer code, and even mimic a student's writing style. The result is a confusing gray area where it is no longer obvious whether a student is using a tool to learn or using a tool to do the work for them. This shift has forced universities to ask a difficult question: when a student uses these powerful machines, are they violating academic integrity, or are they simply adapting to a new reality? The answer, it turns out, is not found in the software itself, but in the human mind behind the keyboard.

A recent study published in the proceedings of the 11th International Academic Studies Congress seeks to understand the psychological forces that drive students to make these choices. Rather than treating academic dishonesty as a simple matter of breaking rules or getting caught by detection software, the researchers, Ezgi Dagtekin and Ercan Erkalkan, propose that the decision to use artificial intelligence honestly or dishonestly is a complex mental process. They examined a collection of sixteen key publications from 2022 to early 2026, looking for patterns in how students think about these tools. Their work suggests that students do not view all uses of artificial intelligence as a violation of academic integrity. Instead, they weigh a variety of internal and external factors, such as how much pressure they feel, what their friends are doing, and how clearly their teachers have explained the rules.

The researchers found that the risk of dishonest behavior rises when the environment feels uncertain. When a university's policies are vague, or when a student believes that everyone else is using these tools without getting in trouble, the psychological barrier to dishonest behavior lowers. In these situations, students often use a mental trick called moral disengagement. This is a way of convincing oneself that a questionable action is actually harmless or necessary. A student might tell themselves that using an artificial intelligence to write a draft is just a form of efficiency, or that it is simply a normal adaptation to a new technological world, rather than a violation of academic integrity. The study highlights that this rationalization is often driven by social influence; if a student perceives that their peers are using these tools to get ahead, they feel a stronger pull to do the same, regardless of the official rules.

Another major factor is the clarity of the instructions given to students. The authors found that broad, abstract rules like "do not use artificial intelligence inappropriately" are often too confusing to be effective. Students struggle to know exactly what counts as acceptable help and what crosses the line into dishonest behavior. When guidance is specific to a particular assignment, such as clearly stating whether it is okay to use the tool for brainstorming but not for writing the final draft, students are less likely to slip into dishonest behavior. The study suggests that the confusion often stems from a mismatch between what a university says in its general handbook and what a specific teacher expects in a specific class. This ambiguity can act as a silent permission slip, leading students to make choices they might not have made if the expectations were crystal clear.

The pressure of the academic environment also plays a critical role. When students face tight deadlines, heavy workloads, or intense anxiety about their grades, they are more likely to turn to artificial intelligence as a way to cope. In these moments, the use of the tool is often driven by a sense of necessity rather than a deliberate intent to deceive. If a student feels they cannot complete a difficult task on their own, or if they lack confidence in their ability to organize their thoughts, they may delegate the hard cognitive work to the machine. This is not always a calculated act of fraud, but rather a response to feeling overwhelmed. The study points out that when students feel they have no other way to succeed, the moral cost of using the tool feels much lower.

To make sense of these findings, the authors propose a model that connects the environment to the student's mind and finally to their actions. Imagine the university setting as a landscape that shapes the weather of a student's decision-making. The rules, the teaching style, and the assessment methods form the landscape. This landscape influences how a student interprets the situation: do they feel pressure? Do they think it is normal to use the tool? Do they believe they are capable of doing the work themselves? These internal thoughts then determine the outcome: the student might use the tool openly and honestly, use it in a way that is on the edge of the rules, or use it to dishonestly replace their own work. The study argues that the environment does not force the behavior directly; instead, it shapes the student's perception of what is right, necessary, and safe.

The implications of this research suggest that simply banning artificial intelligence or relying on software to detect it is not enough. The authors argue that universities need to move beyond punishment and surveillance. Instead, they should focus on creating clear, specific guidelines for every assignment, teaching students how to use these tools ethically, and supporting students who feel overwhelmed by their workload. By addressing the psychological drivers—such as the fear of failure, the confusion over rules, and the pressure to perform—educators can help students make better choices. The study concludes that the solution lies in a combination of clear policy, better teaching, and genuine support, helping students understand that owning their work means taking responsibility for the ideas they submit, regardless of the tools they use to get there.