Analysis

Ivy League faculty are pushing back on AI in public, from Dartmouth's Anthropic deal to Harvard's classrooms

Six Dartmouth professors objected on the record to the college's Anthropic deal, two thirds of Harvard faculty say AI hurt their courses, and a Brown economist's exam average halved.

AI Fin ResearchCovers US
At the top schools the pushback is on the record, with names attached.
130+Dartmouth faculty with work named in the copyright suit against the AI company their college partnered with
2 in 3Harvard faculty who say AI had a negative effect on their classes, up from 42% a year before
96 to 48the final exam average in a Brown economics course once the exam moved in person

Dartmouth signed with Anthropic, and its own professors objected

Dartmouth announced a partnership with Anthropic and Amazon Web Services on 3 December 2025, which The Dartmouth called the first adoption of AI at institutional scale in the Ivy League. The student paper reported on 13 January 2026 that more than 130 Arts and Sciences faculty have publications named in the copyright class action against Anthropic, which the company agreed to settle for 1.5 billion US dollars.

Professor Department On the record
Matthew Garcia Latin American, Latino and Caribbean Studies “They have not honored shared governance”
Bethany Moreton History Anthropic “ripped off the published scholarly work” of the college’s own faculty
Mary Coffey Art History The faculty committee’s only input was that it “helped the College shape the message”
Aden Evens English “Something is deeply wrong when the company that steals without acknowledgement the writing of my colleagues” becomes a partner
Stephen Brooks Government “It’s a big problem that firms that misuse intellectual property face so little consequences”
Darrin McMahon History “I wish there could have been more input from the faculty”

One professor quoted took the other side. Sean Westwood of the government department said: “We could certainly try to pretend that AI is not going to be a requirement for future jobs.”

Dartmouth announces the Anthropic and AWS partnership.Students get Claude for Education.
Six professors criticize the deal in The Dartmouth.Their complaints are about governance and about a partner accused of copying their work.
The provost who announced the deal is accused of having AI write his essays.A detector labeled his Washington Post essay on AI cheating 100% AI-generated. The student paper put his nine professional articles of the year at 96%.
The faculty union chapter petitions for his removal.The president appoints an advisory committee and says: "Leading on something this new and fluid is messy."

The provost, Santiago Schnell, told Inside Higher Ed that the 100% label “is not accurate” and that “I wrote this essay based on my own thinking.” Careless use left a trace a detector could find, and it gave the faculty their strongest argument of the year.

At Harvard, two thirds of faculty say AI hurt their classes

Harvard Magazine reported on 18 September 2026 on the Crimson’s annual survey of the Faculty of Arts and Sciences.

Of 463 faculty who answered Share
Say AI has had a negative effect on their classes Two thirds, up from 42% in 2025
Have received coursework they knew or suspected was AI-generated Nearly 90%

Harvard’s leaders are moving the other way. The president says he is “convinced” AI will speed the advance of knowledge, the dean of the College favors “actively encouraging AI usage where it can deepen learning,” and the College has given its members access to Claude and added a required AI module to first-year writing. The historian Jill Lepore is quoted on the other side: “The machine conversation is not free.”

At Brown, an in-person final cut the average in half

Roberto Serrano, an economist at Brown, suspected AI use on a take-home midterm in spring 2026 and announced an in-person, closed-book final. Many students dropped the course. The final average fell from 96 to 48, GBH News reported. He has since made homework worth nothing toward the grade and added an oral exam by webcam. He is not banning AI. His line: “We cannot choose to become idiots.”

A search of the other five schools turned up no faculty letter

Searches of the student papers at Princeton, Yale, Columbia, Cornell and Penn turned up no faculty letter or resolution against AI. At Penn, Wharton’s dean took the school all-in in 2024.

The objections are about governance, copied work and cheating

Objection Where Does it bear on a researcher’s tools?
The deal was made without the faculty Dartmouth No. It is about who decides
The partner copied our published work Dartmouth No. It is about a lawsuit
Students use it to avoid learning Harvard, Brown No. It is about assessment
An administrator had it write for him Dartmouth Yes, as a warning: careless use gets caught

None of the professors quoted argues that an agent cannot pull data or estimate a model. A finance researcher asking a committee for agent access will meet these arguments anyway. Answer the one that is being made, which is usually about process, and bring a pipeline whose checks the committee can read. The SEC filings guide is one to show them.

Sources

Related

Analysis

Throw out what you knew about AI before September

Fable 5.1, GPT-6 Astra and Opus 5.5 arrived inside five weeks. One science benchmark doubled in two months. A test you ran in the spring describes a different technology.

Anthropic, Introducing Claude Opus 5.5Global