Generative AI for All: an Open-Source GenAI Platform for Your Campus Community

Researchers have measured generative AI productivity gains and job performance improvements ranging from 14% to 40% for those who use it. In the past 12 months, generative AI use has exploded across all job types, with a 62% increase in use overall. By 2030, nearly all industries and job roles will be impacted by generative AI?most notably health care, banking, technology, education, media, and entertainment. To prepare our students for this future and to reap operations and productivity gains for our faculty, staff, and researchers today, Vanderbilt has built and deployed a private generative AI platform named Amplify with features such as Retrieval Augmented Generation (RAG), Custom Assistants, Code-Interpreter support, and Sharing, with easy user access thanks to SSO integration. Amplify supports our equity goals, our privacy and security requirements, is cost-effective, and supports LLMs from multiple providers. During the nine-month pilot across 1,200 users, we averaged $2/user/month in token costs. In this session, we will cover the drivers behind our design approach, how we engaged campus stakeholders to enhance the software, outcomes and lessons learned from our pilot and first four months of production usage, and details of the costs incurred. We’ll close with a brief demo of the platform and welcome institutions to download the open-source software and join the growing community of practice.