Aug. 30, 2023

     We hope you’ve had a restorative summer, and that you’re looking forward to the fall semester here at Princeton as much as the three of us.

     Since entering the daily news cycle last winter, powerful new generative AI tools (like ChatGPT and Bard) have been the subject of sincere excitement and genuine concern within higher education.  These new tools promise to galvanize learning as much as they seem poised to threaten it, and they’ve already begun to shape how the Princeton community reads, writes, codes, conducts research, and accesses academic support.

     In response, we‘re writing to clarify the University’s current approach to generative AI in our courses.  This summer, a working group on generative AI organized by the Office of the Dean of the College met to sort through the literature and to make recommendations to Princeton’s administration about how to address this new tool in teaching and learning.  (The working group, co-chaired by Kate Stanton and Cecily Swanson, included members listed below our signatures.)

     Prof. Meredith Martin, who directs our Center for Digital Humanities, and Prof. Arvind Narayanan, who directs our Center for Information Technology Policy, spoke with the working group.  The group’s report is now being reviewed by the Provost.  In the meantime, we write to share these recommendations.

     First and foremost, the decision to allow, limit, or prohibit generative AI in a course or in undergraduate independent work will remain our faculty’s.  Faculty members have the discretion to set their own generative AI policy for their courses.  We encourage each instructor to establish a clear policy and post it on their syllabus (please see the McGraw Center’s website for more pedagogical guidance).

     Our existing academic integrity regulations govern how students may use generative AI. As you know, Princeton requires students to state the work they submit in a course is original and only their own (RRR 2.4.3).  Just as students may not turn in someone else’s work as their own, students may not misrepresent as their own work any output generated by or derived from generative AI.

     Students who’ve received permission to use generative AI must follow the requirements for acknowledging sources in academic work (outlined in RRR 2.4.6) and must adhere to the scholarly standards for an explicit statement of method.  Students who don’t follow these regulations will be subject to disciplinary action.

     We want to also clarify that, following our current policies about outside tutoring (RRR 2.4.5), undergraduates may not use AI tutoring bots or any tutoring service that isn’t authorized by the Office of the Dean of the College. 

     We believe that detection and surveillance tools are not an effective means to identify or deter the use of generative AI.  They’re not reliable, and they appear to be biased.  We don’t recommend that faculty use these tools.

     Instead, we hope that instructors and advisers will engage students in active conversation about the importance of academic integrity, original work, and source acknowledgment to Princeton’s academic mission.

     We believe that the powers and risks of generative AI should only deepen the University’s commitment to a liberal arts education and the insistence on critical thinking it provides.  We will continue to assess the development of generative AI and will generate more comprehensive guidance as the year progresses.

     We’d be grateful for your input.  If you’re a student, how are you using generative AI in ways that enhance your education without jeopardizing academic integrity?  If you’re an instructor, how does generative AI provide opportunities for engagement with course material in positive and productive ways that also disallow academic integrity infractions?

     Do let us know.

     With our best wishes for your teaching and learning this semester,

            Deans Dolan, Stanton, and Swanson

NB:  Members of the working group on generative AI included Anne Caswell-Klein, Dean of New College West; Joyce Chen, Senior Associate Dean, Office of the Dean of Undergraduate Students; Wind Cowles, Associate Dean, Data, Research and Teaching, Library-Data, Research and Teaching Services, Princeton University Library; James Alec Dun, Associate Dean of the College; Mona Fixdal, McGraw Center for Teaching and Learning; Amanda Irwin Wilkins, Director, Princeton Writing Program; Maria Medvedeva, Director, Academic Engagement, Office of Dean of the College; and Christine Murphy, Associate Dean for Academic Affairs, Office of the Dean of the Graduate School.  We thank them for their time and expertise.