Citizenship in the 21st Century Winning Policy Memo
Introduction
In the 2026 COLLEGE 102 course, members of the class of 2029 collaborated to draft 109 policy memos about a pressing question: How should the Stanford community promote learning, reduce cheating, and conduct accurate assessments of student learning in the age of artificial intelligence?
This memo was written by Neel Ahuja, Erin Chen, Henry (Honglong) Chen, James Chen, Bernardo Dal Molin, Emelina Fry, John Li, Jason Lu, Nathaniel Motulsky, Y-Lan Nguyen, Diego Sanchez, Varin Sikka, Anna Song, Janus Tsen, Ariadne Vidalakis, Peter Vu, and Esme Zeineh (COLLEGE 102 section #38, instructor: Byron Gray). In an anonymous peer review, this was the most-commended memo by a significant margin: of the eight sections that reviewed it, five selected it as their top choice over seven other memos, and two additional sections chose it as their runner-up.
View the assignment Download a printer-friendly version of the memo and assignment
Policy Memo
Artificial Intelligence has outpaced Stanford’s course-level policies, posing a university-wide question of how Stanford can maintain academic integrity while integrating such technology effectively. Since students’ grades depend mostly on take-home evaluations, students may rely exclusively on AI assistance without learning the content themselves (Cotton et al.). Hence, we propose a new form of assessment with restructured grade weighting to ensure student engagement and prepare students to use AI as a learning tool:
Understanding Checks
As a university policy, Stanford should encourage departments to implement brief, periodic in-person assessments—called understanding checks—that supplement existing coursework (Arhavbarien). In lecture-based classes, this could entail oral check-ins or supervised problem-solving sessions. In STEM courses, students might walk through problems with a TA; in humanities courses, students could respond to a prompt or outline an argument under supervised conditions. By evaluating students’ ability to articulate their thought-processes, these assessments measure learning rather than potentially dishonest results (Cheng). This approach lessens the need to police AI use by ensuring that core evaluations occur in settings where understanding is authentically demonstrated.
Weighting
Each short assessment would individually weigh 2–3%, collectively weighing 15–20% of students’ final grades—providing multiple opportunities for students to demonstrate mastery, ensuring equity, and reducing high-stakes pressure (Karpicke and Roediger). Moreover, by verifying understanding through short in-person assessments, students may be encouraged to use AI as a learning aid when working through problems independently. To further emphasize the learning process, departments should incorporate components such as attendance, section participation, and in-class engagement into final grading.
Addressing Concerns
One may worry that a fluke leading to a poor performance could disproportionately impact a student’s grade; however, this is mitigated by the low weight of each individual assessment. Additionally, while one could claim oral assessments unfairly penalize introverted students, conversational skills are essential in professional environments and aid long-term concept retention. To lower anxiety, assessments could be structured as informal conversations. Another potential concern is the increase in instructors’ workload from supervising in-person assessments. However, Stanford can address this issue by redistributing the workload from grading lengthy take-home assignments to understanding checks.
Works Cited
- Arhavbarien, Joseph. “Low-stakes assessments in secondary and further education schools: A systematic literature review.”
- Cheng, Chhayna. “A Review of Benefits, Challenges, and Strategies of Students’ Oral Presentation in Higher Education.”
- Indonesian Educational Research Journal, vol. 3, no. 2, Dec. 2025, pp. 62–84. journal.id-sre.org/index.php/ierj/article/view/113.
- Cotton, Debby R. E., et al. “Chatting and cheating: Ensuring academic integrity in the era of ChatGPT.”
- Innovations in Education and Teaching International, vol. 61, no. 2, 2024, pp. 228–39. doi.org/10.1080/14703297.2023.2190148.
- Karpicke, Jeffrey D., and Henry L. Roediger. “The Critical Importance of Retrieval for Learning.”