Harvard’s Faculty of Arts and Sciences will adopt Anthropic’s Claude Code toolkit while ending access to OpenAI’s ChatGPT Edu after June 2026, according to The Harvard Crimson. FAS Senior Advisor on AI Christopher W. Stubbs confirmed the plan, citing financial considerations and lower-than-expected student usage.
“The uptake among undergraduates was far less than we anticipated,” Stubbs told The Crimson. Students avoided the university-provided ChatGPT accounts because they believed the tool was deployed to catch plagiarism. “That was just emphatically not the case,” Stubbs said. “It was really an attempt to have equity of access across the student population.”
Multi-Platform, Not Single-Vendor
The shift does not represent an exclusive commitment to Anthropic. FAS affiliates will retain access to Google’s Gemini through an existing institutional agreement, and Stubbs indicated the school would continue evaluating tools as the market evolves.
“I think people are going to find that the tools that are available to people with g.harvard.edu authentication are very much functionally similar to what OpenAI is,” Stubbs told The Crimson.
Dean of Engineering and Applied Sciences David C. Parkes acknowledged the transition at a recent event focused on agentic AI in research: “We’ve been providing ChatGPT OpenAI technology, and we will be providing Anthropic technology.” He added that the school wanted students to be “aware of how to use these tools” rather than locked into a single platform.
FAS spokesperson James M. Chisholm said affiliates should not expect a long-term commitment to any single provider. “Given how rapidly evolving the space is, and that you have new technologies and applications and use cases emerging all the time, this is going to be something that is continually evaluated by the FAS,” he told The Crimson.
Faculty Response
Faculty appeared largely unconcerned. Organismic and Evolutionary Biology professor Bence P. Ölveczky wrote that “moving on from ChatGPT is fine with me.” Statistics senior preceptor Kevin A. Rader called losing ChatGPT Edu “a little bit of a bummer” but said the shift would not affect his workflow because he primarily uses a sandbox offering several AI platforms.
The Institutional Signal
The decision is notable less for its direct impact on Harvard operations and more for what it signals about institutional AI procurement. Rather than committing to a single vendor, FAS is maintaining optionality: Claude for agent-capable coding tasks, Gemini through Google’s institutional agreement, and a willingness to switch again if something better emerges. After June 2026, continued ChatGPT Edu access will require administrative and budgetary approval, effectively moving OpenAI from default to opt-in.
For Anthropic, the placement puts Claude Code in front of researchers and graduate students at one of the world’s most visible universities, a distribution channel that builds familiarity among the next generation of AI tool users.