Student ethics and ChatGPT: What college instructors need to know

Artificial intelligence (AI) tools like ChatGPT are prompting educators to face critical questions about what we teach and how to measure student learning in authentic ways. My honors undergraduate student and I decided to start at the beginning—with an examination of how and why students are using AI. We also wanted to look at students’ perceptions of the ethical issues surrounding its use and ask about cheating behavior in general. The resulting study of 569 (primarily first-year) college students revealed important insights about both AI ethics and broader patterns of cheating behavior.

One of our most important findings is that the students demonstrated a clear understanding of what constitutes ethical versus unethical use of ChatGPT in academic settings. Through exploratory factor analysis, we identified three distinct categories of student perceptions about how students might use ChatGPT:

Ethical academic use: activities such as gathering information for assignments, preparing study notes, validating information, and getting background knowledge on course topics were viewed by students as academically appropriate information-gathering activities.

Academic misuse: activities such as answering exam questions, responding to assignment prompts, rewriting others' work, and artificially increasing word count or complexity of responses were viewed by students as unethical.

Non-academic use: “academic-adjacent” activities, such as composing emails or correcting grammar, were viewed by students as acceptable and distinct from academic work.

Our analysis of student motivations revealed two primary factors driving ChatGPT use. Students are primarily motivated by value and convenience—they use ChatGPT because it helps them learn more effectively, provides accessible information, and offers educational value. This contrasts with hedonic motivation, which represents using ChatGPT to "beat the system" or as a shortcut when rushed for time.

The (anonymous) students in our sample reported significantly higher value-based motivations than hedonic ones, suggesting that most are approaching AI tools with genuine educational intent rather than seeking to circumvent learning. This finding suggests that the problem isn't AI itself, but rather students' underlying motivation to learn. When students are genuinely motivated to understand material, they use AI as a learning tool; when they're not, they use it to avoid learning altogether.

However, as the popularity of AI soars, students are experiencing the fear of being left behind when other students in their classes use chatbots to get better grades (The Conversation, 2025); this will likely encourage even greater adoption of AI.

While our findings about ChatGPT ethics are encouraging, the study also revealed sobering statistics about overall academic misconduct that provide important context as we examine the bigger picture. We asked students to indicate, using forced-choice yes/no responses, whether they had engaged in a variety of cheating behaviors during their university studies. The results were disheartening:

• 2% had submitted someone else’s academic work as their own
• 7% had cheated on an exam
• 8% had paraphrased or copied from a source without proper citations
• 4% had used unauthorized electronic resources to complete an assignment
• 3% had collaborated on an assignment when the instructor asked for individual work

Although discouraging, these rates align with previous research showing that academic misconduct is pervasive in higher education, with approximately 75% of students admitting to some form of cheating (McCabe et al., 2015).

Our findings offer several practical insights for psychology educators navigating the AI landscape:

Students need guidance, not prohibition. Since students can distinguish ethical from unethical AI use, instructors can focus on clarifying boundaries rather than blanket bans. Consider explicitly outlining which AI applications and uses are acceptable for your courses and which cross ethical lines (but be prepared to update this—a lot—as AI use evolves).

Emphasize learning over shortcuts. Our research suggests that students' underlying motivation to learn—not access to AI tools—is the key factor determining whether technology enhances or undermines education. Students motivated by genuine learning use AI ethically, so help students understand how to use these tools to enhance rather than replace their thinking. Presenting this new MIT study on the dangers of AI use wouldn’t hurt either.

Address the broader cheating culture. The high rates of traditional academic misconduct suggest that ChatGPT is entering an already problematic ethical environment. In other words, AI may be making it easier to cheat, but the cheating was already happening. Addressing contextual factors for academic misconduct may be more impactful than focusing exclusively on the role of AI.

Model appropriate use. Instructors have an opportunity to shape early adoption patterns by demonstrating appropriate use. For example, Baidoo-Anu & Ansah (2023) discuss various academic benefits of AI, such as personalized tutoring and adaptive learning. Providing students with examples of positive AI use can help to foster a learning-focused educational environment.

Rather than viewing ChatGPT as primarily a threat to academic integrity, our research suggests the real challenge—and opportunity—lies in fostering genuine motivation to learn. Students who want to understand material will use AI as a powerful learning tool; those who don't will find ways to cheat (with or without it). Our role as educators is to cultivate that intrinsic motivation while providing clear guidance on how AI can enhance rather than replace learning.

Jenel Cavazos, PhD, received her PhD in social psychology from the University of Oklahoma where she is an associate professor of psychology and introductory psychology coordinator. Her primary role in the department of psychology is to teach two large sections of PSY 1113 each semester and to supervise the graduate students and adjuncts teaching several other sections. She is responsible for curriculum development as well as providing mentorship for graduate and undergraduate teaching assistants. She also serves on several university committees related to general education, learning technology, and student success. Her primary research interests include the scholarship of teaching and learning and teaching with technology.

Keane Hauck recently received his BS in psychology from the University of Oklahoma with minors in computer science and music. He plans to pursue a PhD at Arizona State University in the fall of 2025, where he will study quantitative psychology under the mentorship of Kevin Grimm. In his free time, Keane competes in local tournaments for the Pokémon trading card game.