Harvard makes AI training mandatory for all first-year writers
Harvard College has made a three-part AI training module mandatory for freshmen in its expository writing program this semester. More than 90 percent of Harvard's incoming class takes an Expos course to fulfill the College's writing requirement, making this the broadest curriculum move on generative AI at a major U.S. university to date.
The module covers how large language models work, their effects on education, and broader questions including environmental impact, data privacy, copyright, and disinformation. The Harvard College Writing Program developed the training after piloting it in select courses last fall.
A common foundation, not an agenda
The goal was to give every student baseline knowledge about AI and reduce pressure on faculty to teach the mechanics themselves. Gillian B. Pierce, associate dean of undergraduate education, told The Harvard Crimson: "Faculty may have some sense that, 'Oh, I need to teach students what is an LLM.' You can kind of assume now that students will come knowing that."
Writing Center director Jane Rosenzweig and preceptor Tad Davies, who led the module's design, said the training was meant to equip students to engage with ongoing debates rather than steer them toward any particular view of the technology. Rosenzweig said: "There's no agenda for what students should take away from this, except what we all agree on, which is that when you have information, you can make better judgments about what you want to do with the technology."
What the module covers
The first section walks students through how large language models function. The second examines AI's effects on education, drawing on Harvard faculty research including studies on custom chatbots in physics courses and the AI assistant used in Computer Science 50.
The third section addresses environmental impact, data privacy, copyright, and disinformation. It includes a close reading of the terms of service for popular AI chatbots-practical work for writers who may use these tools professionally.
Built to evolve
Rosenzweig told The Crimson the module is intentionally iterative. The team re-tests lessons every semester because outputs from image generators and chatbots used as teaching examples change as underlying models update. "This is a work in progress," she said. "As society's conversation about AI changes, expect the module to keep up with that as much as possible."
Davies framed the module as a foundation rather than a complete treatment. Discipline-specific applications are expected to follow in later coursework. "There's going to be disciplinary differences in the way that the models can engage in certain kinds of materials very effectively and less effectively in others," he said.
Pierce positioned the Expos module as a first step toward a broader university response. The module has been shared with the Dean of the College, the Dean of Undergraduate Education, and the FAS Faculty Advisory Committee on Artificial Intelligence.
Harvard administrators have made AI curriculum a priority this academic year. The Dean of Undergraduate Education hosted a Generative AI summit in January, and the Dean of the College has publicly urged students to learn to use AI while warning against letting it displace core learning.
For writers looking to build skills in this area, resources on prompt engineering and AI for writers offer practical training beyond the foundational knowledge Harvard is now providing to all freshmen.
Your membership also unlocks:
