GenAI & Academic Honor: Your Fall Term Playbook

Last week, DCAL and LDI hosted two conversations on the use of GenAI in the academic setting as our community prepares for the start of fall term. We hope these takeaways will help you as you finalize your plans for the weeks ahead.

Navigating Academic Honor & GenAI

The first, Navigating Academic Honor in the Age of GenAI: What We've Learned So Far, addressed our collective need to reorient towards our institutional commitment to academic integrity in the context of GenAI. Katharine Maguire, Director of the Office of Community Standards and Accountability, offered observations from the past year and fielded questions from instructors about the current state of academic honor in our community. The discussion included strategies for crafting more effective GenAI course policies, practices, and expectations with students.

Relevant to this discussion are the institutional Academic Honor Principle, Arts & Sciences Honor Policy, and recently updated Guidelines on use of GAI in Coursework. While these policies aim to provide broad guidance across the institutional context, they operate at a high level only, leaving room for flexibility and necessitating interpretation within specific settings. This need for interpretation–for individual instructors and students within particular courses–is both a challenge and an opportunity.  

As the institutional channel for allegations and reports of undergraduate student misconduct, the Office of Community Standards & Accountability is a key partner in helping the Arts & Sciences community navigate the intersection of GenAI, these policies, and our learning environments. Maguire shared that in 2024-25, the Committee on Standards (the panel within the Office that hears misconduct cases) aimed to lower the barriers to reporting for instructors by piloting a wider range of disciplinary outcomes. In addition to the previously standard two-term suspension, the Committee added a deferred suspension (a probationary period during which a suspension is not enacted until a second violation) and an administrative hearing, during which the reporting instructor and the student work together to determine a resolution. These changes in procedure, alongside the increasing ubiquity of GenAI, resulted in an increase from an average 35 reported academic integrity cases to 75 reported cases in 2024-25. Maguire sees this increase as an improvement in our community's ability to navigate the changing landscape of academic integrity.

Meanwhile, Maguire observed that the motivation for student cheating has not changed: students cheat when they doubt their ability to successfully complete assigned work. This doubt commonly stems from perceived time pressures or a view of assignments as boxes to check rather than important parts of their learning process.

Based on these observations, Maguire offered this advice on navigating academic honor in the current educational moment:

• Talk with students about academic integrity–why it matters for students to engage in their own learning, even (and perhaps especially) when doing so is a challenge. Students need to understand that much of the value in learning comes through struggle, errors, and perseverance. Discussing how the process of learning is different from the process of completing assignments can be helpful. Help students understand why a "bad" grade is a better choice than an 'A' earned through academic dishonesty.
• Engage with students about their GenAI use frequently, and specifically if you suspect that their use falls outside your course policies or expectations. Giving students the chance to engage in a moment of reflection about their use of technology, how it has affected their experience, and what they might be hoping for in the future is critical to our community's ability to uphold academic honor in this changing context. 
• Consider who is getting "caught" through our disciplinary systems, namely students using the free models of GenAI tools rather than the more sophisticated and less detectable paid models. Exploring all the tools available to the Dartmouth community via chat.dartmouth.edu with your students is one way of encouraging greater equity.
• When setting policies and expectations in your courses, grapple with how rapidly GenAI capabilities are evolving, thus making detection increasingly difficult.
• But what about AI detection tools? The prevailing opinion among the educational technology community is that the risks associated with false positives of GenAI detection, along with the high rate of false negatives, outweigh the tools' usefulness. Further, running student work through an AI detection tool without their permission represents a violation of Dartmouth copyright policy and, potentially, FERPA.  It is possible that the effectiveness of these tools will improve over time and this recommendation may change (Learning Design and Innovation is keeping close tabs on this landscape). In the meantime, consider these recommendations for cautious exploration.
• Reach out to the Office of Community Standards if/when you are unsure how to proceed with student interactions around academic integrity and/or GenAI. The Office can take into account any relevant background information or past interactions with particular students that can helpfully inform next steps.

Policies and Practices That Work

The second session, Setting GenAI Policy for Your Course, featured five faculty panelists from a wide variety of disciplines speaking about their varying approaches to GenAI course policies. Panelists included:

• Lucas Dwiel, Psychological and Brain Sciences and Psychiatry
• Tim Pierson, Computer Science
• Tiina Rosenqvist, Society of Fellows, Philosophy
• Chris Sneddon, Geography and Environmental Studies
• Miya Xia, Asian Societies, Cultures, and Languages

In degrees of openness, these instructors ranged from fully embracing student use of GenAI for course-related work to setting limits, banning it entirely, banning it in specific situations, and choosing not to set a policy altogether. Several panelists alluded to changes in their thinking and policies on GenAI use based on experience, and trial and error, in previous teaching terms. They discussed the particular learning objectives of their courses and the disciplines within which they are situated, both of which influenced their decision-making in relation to GenAI course policies.

A common practice among the panelists was regularly discussing GenAI with their students, addressing topics including: 

• The instructors' own perspectives about GenAI
• Ways in which GenAI is being used and perceived in their respective fields
• Tasks for which GenAI is and is not well-suited in course and disciplinary contexts
• The affordances and deficiencies of these tools
• The ethical considerations of their development and use
• How GenAI influences the learning process as well as our fundamental ideas about learning 

In sharing about these student conversations, the panelists reflected a curiosity and willingness to develop their ideas with their students, approaching their course policies and practices from a place of collective experimentation. Several acknowledged the effort required to navigate these open questions and maintain an experimental mindset, as well as the additional challenge of doing so in both larger and more introductory-level classes. The panel discussion emphasized that the role of the educator is to help students set clear and context-appropriate boundaries, to facilitate conversations about how and why the boundaries are meaningful, and to open the door for discussion when the boundary is not working.

Several concrete strategies emerged from the discussion:

1. Discuss with students the relevant pitfalls of using GenAI in your particular discipline, particularly for novices. Dig into the prompts that are necessary for a novice to make the most of the tool. Discuss when/why a novice may not find GenAI a useful tutor. Based on your course's learning objectives, determine whether GenAI is appropriate to use for organizing one's thoughts, summarizing a defensible point, helping to write a research paper, or other relevant learning processes. Encourage students to attend office hours to discuss their use of these tools with you.
2. If unassisted writing or recall are important to your course's learning objectives, have students write essays or complete written or oral exams during class time within the classroom (accounting for potential accessibility concerns). Consider scaffolding these assessments by incorporating drafts, practice, feedback, and revision. 
3. Ensure you clearly articulate expectations for proper citation and acknowledgement if/when students use GenAI tools for course-related work.
4. Explore options for policies, over 200 of which are available in this repository. Share yours with colleagues and ask them to share theirs! 
5. Consider how you might augment your course policies with thoughtful assignment design that takes GenAI into account. To this end, join us for the Teaching with Gen AI Institute (next up: December) and/or pursue a Teaching with GenAI grant, both of which panelists noted were instrumental to their progress. 
6. Learn more about the Writing Program's exploration of building AI Literacy skills with students. 
7. Book a consultation to talk with LDI and DCAL about GenAI policies and usage in your class!

Our thanks to Katharine Maguire, Chris Sneddon, Miya Xie, Lucas Dwiel, Tim Pierson, and Tiina Rosenqvist for participating in these sessions and sharing their experiences!