Large Language Models at UBC

The Provosts created a cross-campus Steering Committee to explore the benefits and risks associated with using Generative AI in the university context. Given the broad range of university operations that could potentially be impacted or benefited by Generative AI, this committee created several subcommittees with focus on Teaching and Learning, Large Language Models, and Conversational Tools.

The focus of this subcommittee will be on Large Language Models (LLM). The power of Generative tools such as ChatGPT is primarily due to the underlying LLMs that have the capability to generate, extract and summarize new data/information from previously trained LLMs. While it is possible to use off-the-shelf LLMs to enhance university operations, we are interested in exploring the possibility of creating university specific LLMs that can enhance the student experience and improve the efficiency of the university’s operations by including data that are context specific to UBC. There are some obvious challenges to building LLMs including the significant data and computing resources that are required to build a foundation model, the ethical issues and inherent biases in the generated content, and the privacy and security of data given to and generated by LLMs. Our goal is not to build a foundation model but to identify ways in which we can either embed or extend the existing foundation models to include UBC specific content in a way that is secure and cost effective.

Membership

The membership is comprised of a broad range of faculty, student and staff roles across both campuses with a connection to and providing expertise on various dimensions of generative artificial intelligence and teaching and learning.

Co-chairs

Jennifer Burns
Associate Vice-President, Information Technology and CIO

Bhushan Gopaluni
Vice-Provost and Associate Vice-President, Faculty Planning

Committee Administrator: Amy Kao, Manager, Administrative Support 

Members