Guidance on Preventing and Identifying Unauthorized Artificial Intelligence Usage

According to the Honor Code Charter, “Oberlin College & Conservatory students are on their honor to uphold academic integrity…all work that students submit is expected to be of their own creation and give proper credit to the ideas and work of other individuals and resources” (Oberlin College & Conservatory, 2026, p. 1). It is essential for students to develop and submit their own work to ensure that they meet the educational goals and requirements of the Oberlin curriculum. At the same time, access to generative technologies, such as artificial intelligence (AI), continues to increase, which provides easier access to tools that may be used, whether intentionally or inadvertently, to develop and submit work that is not students’ own. This guide is meant to guide faculty on both preventing unauthorized artificial intelligence usage and identifying unauthorized artificial intelligence usage within coursework. 

Preventing Unauthorized Artificial Intelligence Usage

While faculty may not be able to prevent all acts of academic misconduct, there are some actions that faculty can take to minimize the temptation or applicability of artificial intelligence misuse. Those tips include:

• Develop a Clear Syllabus Statement: Ensure that your syllabus contains a statement relating to the Honor Code and permissions regarding the usage of artificial intelligence (AI). Education Developer, Dr. Lance Eaton, has collected sample syllabi statements for Generative AI tools by discipline. 
• Facilitating a Dialogue about the Importance of Ethics in Academics: Going beyond the syllabus statement when it comes to ethics and artificial intelligence can be beneficial. At the beginning of your course, discussing the implications of AI misuse in the subject and potential career fields relating to the course can help students learn about the importance of submitting their own work and properly attributing resources, including technology (see the resources below, Harvard University, n.d.).
• Assignment Prompts Clarify AI Parameters: Assignment prompts, especially for papers or other written assessments, should include specific parameters for AI usage. Specific prompts can help deter and detect unauthorized tool use (University of Southern California, n.d.).
• Exam Proctoring and Testing Environment: Under the revised Honor Code charter, faculty are now permitted, but not required, to proctor in-person examinations (Oberlin College & Conservatory, 2026). A supportive faculty presence can help serve as a resource to students in the class and deter academic integrity issues. For more guidance on proctoring, please visit Oberlin’s proctoring guidance document.

Tips to Identify Unauthorized Artificial Intelligence Usage

If you suspect unauthorized artificial intelligence use on an assignment, here is some advice as you investigate the matter further:

• Use Empathy: Confronting a student and being confronted with a potential academic integrity issue can be stressful, and the way a potential violation is addressed can influence a student’s response. Students have reported increased stress regarding potential false or wrongful accusations regarding artificial intelligence misuse (Stephenson & Armstrong, 2026). Consider framing your initial inquiry in terms of educational growth, for example, “It’s important for all students to acquire [specific knowledge and skills] in this course, and they do so by practice and by exercising their own judgment. The work I received from you causes me to wonder whether all of this is your own. Let’s talk about how you arrived at these results/insights/conclusions.”
• Assess for Content Competency: During your meeting with the student to discuss the issue, ask questions regarding the course content to assess if they understand the course content. This can help you and the student to identify places where it may have been or may be tempting to reach for answers outside of their own scope of knowledge and skills. 
• Be Transparent: Not disclosing the information that led to your suspicion can often result in the student also hiding information. Addressing the issue with transparency about your suspicion can help influence a more thorough response from the student. Point to any questionable areas of submitted work that do not align with what you have taught or required or that differ unusually from the student’s previously submitted work. Ask the student about their working methods and ask to see versions from their drafting process.
• Use Detection Tools with Caution: AI detection tools are not completely reliable and have shown bias through their responses (Liang et al., 2023). While these tools can help point out certain identifiers of AI usage, their results alone cannot constitute a preponderance of evidence that AI was misused. It is important not to overstate their reliability.
• Check Citations: Often, AI can generate wrong links or hallucinate citations when it comes to academic papers. In fact, ChatGPT leaves a remnant in the URL hyperlink in papers if ChatGPT was used to find a website (utm_source=chatgpt.com). 
• Inconsistencies Between Submissions: Checking for inconsistencies between a student’s writing, either in one assignment or multiple, can help point to potential technology misuse. However, be cautious when confronting improvements in performance through a course, as improvement is of course a desirable outcome of good teaching and good efforts to learn.  
• Google Document Version History: Asking students to write assignments completely in a Google Document can be a helpful way to view version history to assess any suspicious text sections. 
• Handwritten Assignments: whether in part or in full, can help to focus students’ attention and reduce overreliance on electronic resources.  
• Consult with the Honor Committee: If you have questions, either before or after an initial conversation with the student, feel free to consult with the Honor Committee! You may contact either the Faculty Honor Committee (check Blackboard Committees for the most up-to-date membership), Student Honor Committee (ohonor@oberlin.edu), or the Honor System Liaison (Thom Julian; tjulian@oberlin.edu). 

Several other resources can be helpful to you as you further develop your pedagogical practices to support academic integrity in the ever-changing context of AI innovation. Here are just a few links that can help you through this process:

• East Central College: Detecting AI-Generated Text: Things to Watch For
• Harvard University: Academic Integrity and Teaching With(out) AI
• MIT: AI Detectors Don’t Work. Here’s What to Do Instead.
• University of Massachusetts Amherst: How Do I Address Suspected Student Misuse of GenAI Tools?
• University of Nevada, Reno: How can I revise my assignments to deter student use of AI?
• University of Southern California: Addressing Generative AI

References

Harvard University. (n.d.). Academic integrity and teaching with(out) AI. Office of Academic Integrity and Student Conduct. https://oaisc.fas.harvard.edu/academic-integrity-and-teaching-without-ai/

Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). GPT detectors are biased against non-native English writers. Stanford University, 1-9. https://doi.org/10.48550/arXiv.2304.02819 

Oberlin College & Conservatory. (2026). Honor Code Charter. https://www.oberlin.edu/dean-of-students/student-conduct/academic-integrity

Stephenson, R., & Armstrong, C. (2026). Student generative artificial intelligence survey 2026 (HEPI Report 199). Higher Education Policy Institute. https://www.hepi.ac.uk/wp-content/uploads/2026/03/HEPI-Report-199-Gen-AI-Survey-2026.pdf

University of Southern California. (n.d.). Addressing generative AI. Office of Academic Integrity. https://academicintegrity.usc.edu/faculty-resources/addressing-generative-ai/