Key Takeaways
- Only 18 of Harvard College’s 50 concentrations have department‑wide policies governing AI use in senior theses; the remainder leave the decision to individual students and advisers or are still drafting rules.
- Six concentrations ban AI at every stage of the thesis process, while ten permit AI for research or data analysis but prohibit its use in writing the thesis itself.
- Bans are most common in the Arts and Humanities and Social Sciences divisions; no concentration in the Science Division or the School of Engineering and Applied Sciences (SEAS) has instituted a total prohibition.
- Departments that allow limited AI use typically require students to disclose how the technology was employed (e.g., Economics mandates an appendix detailing AI‑assisted work).
- Faculty and students express divergent views: some see AI as a threat to independent thinking, while others argue that restrictive policies hinder preparation for professional work where AI is routine.
- Dean of Undergraduate Education Amanda Claybaugh advocates discipline‑specific policies that mirror how professionals in each field use AI, emphasizing that the senior thesis should reflect authentic scholarly practice.
Harvard’s Patchwork of AI Policies for Senior Theses
A Crimson review of departmental websites, Canvas pages, handouts, and interviews with more than a dozen department leaders and senior concentrators found that only 18 of the College’s 50 concentrations have established concentration‑wide guidelines for AI usage in theses. The remaining 32 either defer to individual thesis advisers, have yet to finalize a policy, or declined to comment. This fragmented landscape mirrors the broader uncertainty across academia about how generative AI should be integrated into undergraduate research while preserving academic integrity.
Where AI Is Completely Prohibited
Six concentrations enforce a total ban on AI at all stages of the thesis process—from initial research to final writing. These bans are clustered in the Arts and Humanities and Social Sciences divisions. Notably, Government, Social Studies, and Studies of Women, Gender, and Sexuality (WGS) all prohibit any AI involvement. The Social Studies policy, which took effect this year, outlines possible penalties ranging from grade deductions to referral to the Honor Council. In a Canvas announcement, Social Studies 60 instructor Charles Clavey warned students that “a number of papers were written either partially or entirely by generative AI… This is extremely disappointing; it is a disservice to yourselves and disrespectful to your peers and to us.” He added, “Using generative AI makes a mockery of our commitment to teaching and mentorship and the spirit of academic inquiry.”
Limited AI Allowance for Research and Analysis
Ten concentrations permit AI for tasks such as literature reviews, data analysis, or coding but forbid its use in drafting the written thesis. Economics exemplifies this approach: students may employ AI tools for research and analysis, yet they must disclose in an appendix exactly how the technology was used. Jeffrey A. Miron, Director of Undergraduate Studies for Economics, explained, “The crucial thing that we are sort of trying to emphasize in our policy is to be forthcoming, to be clear about what part was done with AI and what part not done.” He added that while AI can assist with literature reviews and coding, “the crucial question in a huge fraction of undergraduate theses… is whether you can address correlation versus causation, and AI doesn’t really help you with that.”
Arts and Humanities: Predominantly Pending Policies
Within the Arts and Humanities Division, 13 of 18 concentrations have not yet settled on a formal AI policy. Only three—History and Literature, Philosophy, and Slavic Languages and Literatures—have instituted total bans. Philosophy concentrator Sofia E. Mikulasek ’27 captured the rationale behind these prohibitions: “Using AI is not thinking… That needs to be something that you yourself do.” Conversely, many humanities concentrators, such as joint Comparative Literature and Germanic Languages and Literatures student Kyler Hoogendoorn‑Ecker ’27, argue that small departments should create clear, formal guidelines rather than rely on ad‑hoc adviser negotiations, which he deems “not a sustainable system at all.” Germanic Languages and Literatures professor John T. Hamilton noted that his department is “currently working on formulating an explicit policy, which we will promptly publish on our department website and through other pertinent avenues.”
Social Sciences: A Split Approach
The Social Sciences Division displays the sharpest divergence. While Government, Social Studies, and WGS enforce total bans, Economics allows limited AI use with mandatory disclosure. Social Studies concentrator Sydney C. Black ’27 supports the strict stance, saying, “I think leaning towards a more restrictive view is appropriate because we simply just do not know enough yet about how artificial intelligence will affect the way we think… The evidence we do have suggests it’ll be extremely bad.” In contrast, fellow concentrator Vikram M. Kolli ’27 finds the zero‑tolerance rule frustrating, contending that it “prevents students from succeeding when they graduate college.” The Social Studies policy also permits students to apply for exceptions, which must be approved by department leadership; WGS grants exceptions only in limited data‑analysis cases, while Social Studies evaluates each request case‑by‑case.
Sciences and Engineering: Adviser Discretion Predominates
Most concentrations in the Science Division leave AI usage to the discretion of individual thesis advisers. This applies to all Engineering students pursuing the S.B. degree, who enroll in the same thesis course, ENG‑SCI 100HF, where AI use is permitted at the adviser’s discretion. Some SEAS concentrations are going further by preparing students to employ AI in research. Applied Mathematics, for instance, advises thesis writers modeling‑based projects to take Applied Mathematics 115: “Mathematical Modeling,” which requires students to use AI in coursework. The concentration’s Canvas page notes, “In Fall 2026, thesis students who are pursuing a thesis that involves modeling should strongly consider taking APMTH 115 to gain knowledge of how generative AI is used in mathematical modeling.”
Integrative Biology exemplifies the rationale behind the adviser‑discretion model. Assistant Head Tutor Andrew J. Berry wrote, “IB doesn’t have a published policy because we expect students and their direct mentors to address these issues as appropriate within the broad guidelines of the Honor Code… IB students write theses in a wide range of biological sub-disciplines, and AI practices may differ from one area to another.”
Dean of Undergraduate Education’s Perspective
Dean Amanda Claybaugh argued that each discipline should craft its own AI policy, reflecting how professionals in that field utilize the technology. She stated, “My own thought… is that the senior thesis is the first opportunity for students to do original research in a discipline, and so the policy governing the thesis should reflect the appropriate use of AI in that discipline.” Claybaugh added that if senior researchers routinely use AI for data analysis, then undergraduates should be allowed to do the same in their theses. Her stance underscores the belief that policy should be aligned with authentic scholarly practice rather than imposed uniformly across disparate fields.
Challenges, Enforcement, and the Road Ahead
Even where policies exist, enforcement remains uneven. Some departments, like Social Studies, have spelled out penalties for violations, while others rely on the Honor Code and adviser oversight. The rapid evolution of AI tools means that today’s allowances may quickly become outdated, prompting departments such as Economics to anticipate policy revisions as the technology advances. Meanwhile, students continue to voice concerns about both the stifling effect of bans and the potential erosion of independent thinking if AI is overused. As Harvard navigates this terrain, the balance between fostering innovation and preserving the rigor of undergraduate research will likely remain a central topic of faculty discourse and student debate.
This article draws on publicly available departmental materials, interviews with faculty and student concentrators, and statements from Dean Amanda Claybaugh. Quoted material is reproduced verbatim from the original Crimson reporting.
https://www.thecrimson.com/article/2026/9/25/harvard-thesis-ai-policies/