Dartmouth Chat is a web-based conversational AI platform developed by Dartmouth's Research Computing and Data Services in partnership with Dartmouth Libraries and hosted at the Massachusetts Green High Performance Computing Center (MGHPCC). Available free of charge to Dartmouth students, faculty, and staff, the platform provides access to a range of large language models within a Dartmouth-managed environment. Users can work with both open-source models hosted locally and leading commercial models from providers such as OpenAI, Anthropic, and Google, all through a single interface and subject to a daily usage limit.

Designed to make generative AI more accessible across campus, Dartmouth Chat gives researchers, educators, and students an institutional alternative to public-facing AI tools while lowering barriers to experimentation and adoption. The platform is already supporting research across a variety of disciplines. In one study, researchers examine how students interact with AI during clinical reasoning tasks, using prompting behavior as a measure of self-regulated learning. Another project draws on ophthalmology literature to generate synthetic medical records for training and analysis. A third investigates the energy demands of campus computing and helps communicate those impacts as part of a broader effort to promote transparency and sustainability in AI. Across teaching, clinical research, and sustainability initiatives, Dartmouth Chat is becoming an increasingly important part of how computational work is carried out at Dartmouth.