TLDR: The Harvard Crimson published an opinion piece arguing that Harvard University’s current generative AI policy is inequitable. The author, Salma O. Siddiqui, highlights disparities in AI usage among students, with men, higher-income, younger, and more educated individuals being more likely to use these tools. The policy, which leaves AI regulation to individual educators, is criticized for creating an uneven playing field and potentially rewarding those who violate prohibitions without getting caught. The article advocates for universal AI literacy training and a more inclusive approach to integrating AI into the curriculum to ensure all students are prepared for an AI-driven world.
The article, written by Salma O. Siddiqui, a Crimson Editorial editor, discusses the perceived inequities in Harvard University’s generative AI policy. Siddiqui notes that many course Canvas pages state that AI use is not permitted and will result in a failing grade. However, she argues that generative AI will inevitably be used by students and could enhance learning if properly integrated.
A study from the Harvard Kennedy School revealed significant disparities in AI usage: men are 7% more likely than women to use generative AI at home, a number that rises to 9% in the workplace. The study also found that higher-income, younger, and more educated workers use generative AI more frequently and effectively. Siddiqui observes this trend among her Harvard classmates, noting that those from high schools with “AI literacy” training tend to use tools like ChatGPT more effectively for tasks like rephrasing or generating counterarguments, rather than simply writing essays.
The current policy has led to 30% of Harvard students worrying about peers gaining unfair advantages through generative AI, according to a Harvard Undergraduate Association survey. Despite this, female students and students of color are observed to avoid the tool even when permitted.
Siddiqui asserts that generative AI, like calculators and spell check, will become a standard workplace tool, necessitating that all students learn to use it effectively. She points out that Harvard’s AI Sandbox offers a secure environment for AI use, yet the official University policy delegates AI regulations to individual educators, which she deems problematic.
A 2023 survey indicated that 54% of college students consider AI use on assignments academically dishonest, but 56% still engage in it. Harvard discourages the use of unreliable generative AI detectors. The author suggests that in courses prohibiting AI, students who use it gain advantages like reduced stress and more sleep without being caught, effectively being rewarded for rule-breaking.
Also Read:
- Generative AI Prompts US College Students to Re-evaluate Higher Education’s Worth
- AI Observatory Unveils Framework to Bridge Global Learning Disparities with Artificial Intelligence
The article concludes by citing research from 2025 showing that 67% of college students believe AI use is “essential,” yet only a third receive AI training from their institutions. Siddiqui references new College Dean David J. Deming’s convocation address to the Class of 2029, where he stated their education would prepare them for an AI-driven world. She argues that under the current policy, this promise is not true for all students, emphasizing the need for Harvard to prepare everyone, regardless of background, for an AI-transformed future.
