Three Universities Show Different Paths Forward on AI Policy
Three liberal arts and public universities have taken distinct approaches to generative AI and LLM use on campus, ranging from structured mandates to decentralized autonomy. Emporia State University implemented a state-mandated policy. Oberlin College launched a year-long institutional examination. Swarthmore College opted against a college-wide policy, instead issuing guidelines and letting faculty set their own rules.
Emporia State: Structured Framework From the Top
Emporia State introduced an AI policy last year based on guidelines from the Kansas Board of Regents and the state's Office of Information and Technology Services. The policy covers ethical use, academic integrity, data privacy, and security.
The university's AI Task Force completed its work earlier this year, having developed training programs and resources for students and faculty. A new requirement takes effect in Summer 2026: faculty must address AI use in their syllabi.
Students access AI training and guidance through Canvas, the university's learning management system. Faculty have their own dedicated resource section.
Oberlin: Institutional Exploration With Mixed Results
Oberlin designated this academic year as a "Year of AI Exploration," supported by the president's office and divisional deans. The college held panels, launched a Critical AI Studies Minor, and gave faculty and staff access to premium versions of ChatGPT and Gemini.
The response from campus has been uneven. Some professors embraced the technology in assignments and classroom instruction. Others reverted to traditional methods like bluebook exams, citing uncertainty about how to handle AI in their courses.
Student reaction split sharply. In the fall, the Luddite Club became one of the most vocal groups against AI use, publishing a letter to the president on a typewriter. Despite the opposition, Oberlin continued forward, awarding AI micro grants to faculty in April.
Swarthmore: Decentralized Decision-Making
Swarthmore has not imposed a school-wide policy on generative AI use. Instead, the college's Information Technology Services published guidelines on privacy, purchasing, and data security.
The rules prohibit sharing "non-public, confidential, proprietary, or otherwise sensitive information to any AI tool" without a specific contract between the college and the service provider. All AI tool purchases using college funding require security approval from ITS.
The Provost recently convened a committee of faculty and staff to guide future policy changes. This approach reflects faculty preference: roughly half of respondents said they would permit but limit AI in their classes, while a third favored an outright ban. Several faculty expressed appreciation for the ability to set policies at the departmental or personal level.
Student usage varies widely. A 2025 survey found 55% of students use AI for academic work once a month or less, 28% use it weekly, and 15% use it daily.
Swarthmore has offered AI for Education courses since 2022, when it participated in the National Humanities Council's Responsible Artificial Intelligence Curriculum Design Project. The college now offers courses on artificial intelligence, ethics, and politics, as well as human-AI interaction.
President Val Smith emphasized the role of the liberal arts in an age of AI. "This is an opportunity to speak powerfully about the importance of the arts and humanities at a time when we have seen both a striking emphasis in terms of technology," she said. Smith stressed the need to educate students in "ethical values, and creativity, collaboration, communication" - skills that extend beyond technical competency.
Your membership also unlocks:
