Measuring University Students' Ethical Judgments of Generative AI Use: Development and Initial Validation of Four Instruments for Academic Integrity Research
This study develops and validates four new instruments for measuring university students' ethical judgments regarding generative AI, revealing that perceived institutional integrity is the strongest predictor of these judgments and that students distinguish sharply between unauthorized AI use in assignments versus exams.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the modern classroom, a new kind of tool has arrived, one that can write essays, solve problems, and generate ideas with startling speed. This technology, known as generative artificial intelligence, has changed the landscape of learning, but it has also blurred the lines of what counts as honest work. For decades, educators have relied on the idea that a student's decision to violate academic standards is shaped by their own character and the rules of their school. We know that when a university makes its policies clear and enforces them consistently, students are less likely to break the rules. We also know that a student's personal drive to learn and their confidence in their own abilities play a role in how they behave. However, until now, researchers have lacked a reliable way to measure how students actually judge these new, gray-area behaviors. Without a clear map of student thinking, it is difficult to understand why some students see using a chatbot to edit a paragraph as acceptable, while others view it as a serious offense.
A team of researchers at the University of Kansas and the University of Texas at San Antonio set out to build that map. They recognized that to understand the future of academic integrity, they first needed better tools to measure it. Instead of relying on old surveys designed for a different era, they created and tested four new ways to ask students about their thoughts. They wanted to know how students rate the seriousness of using artificial intelligence, how they perceive their school's rules and enforcement, how confident they feel using these tools for schoolwork, and what drives them to learn in the first place. By gathering data from 149 students at a large public university, the researchers were able to see if these new questions actually worked and what they revealed about the student mind.
The researchers began by crafting a survey that asked students to rate specific scenarios. They took a classic method used for decades to measure violations of academic standards—asking people to say how serious they think a behavior is—and updated it for the age of artificial intelligence. They included clear-cut violations, such as submitting a paper written entirely by a computer, alongside more ambiguous situations, like using a tool to help brainstorm ideas or to rephrase a sentence without credit. They also built new questions to measure how students view their institution's integrity climate. This concept refers to the student's perception of whether the school has clear rules, whether teachers communicate those rules well, and whether the school actually enforces them. Finally, they added questions to gauge a student's confidence in using these tools for academic tasks and their underlying motivation for being in school.
When the researchers analyzed the answers, they found that their new tools worked exactly as intended. The questions about ethical judgment, institutional integrity, confidence, and motivation each measured a distinct part of the student experience. They were not just different ways of asking the same thing; they captured unique aspects of how students think. The data showed that students are capable of making fine distinctions. They did not treat all uses of artificial intelligence as the same. For instance, students rated using a tool to help during a test as a very serious form of violation, but they rated using the same tool to help with a regular assignment as significantly less serious. This suggests that students are not simply ignoring the rules; they are actively weighing the context of the situation.
The most powerful finding, however, concerned the role of the school itself. The researchers discovered that a student's judgment of what is right or wrong was most strongly linked to how they perceived their institution's integrity. When students felt that their university had clear guidelines, that faculty communicated those guidelines consistently, and that the school enforced the rules fairly, the students tended to view the use of artificial intelligence more strictly. In other words, the environment created by the school mattered more than the student's personal confidence in using the technology or their general motivation to learn. While a student's confidence in using artificial intelligence and their personal drive to learn were interesting factors, they did not add much to the explanation once the school's climate was taken into account.
This study provides a solid foundation for future research by offering a set of validated tools that other scientists can use to study this rapidly changing field. It suggests that the path forward for universities is not just to ban or allow specific technologies, but to focus on the clarity and consistency of their messages. If a school wants its students to make ethical choices, the research indicates that the most effective approach is to ensure that the expectations are clear, the communication is consistent, and the enforcement is credible. The students are watching these signals, and their judgments are shaped by what they see. By understanding this dynamic, educators can move beyond simple detection and toward a more nuanced conversation about how to use these powerful tools responsibly.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.