TerrierGPT Provides BU Community with Free Access to Leading Chatbots
New tool is the result of a partnership between the University’s Artificial Intelligence Development Accelerator and Information Services & Technology
TerrierGPT, a new offering from Boston University’s AI Development Accelerator and IS&T, provides BU community members with free access to leading AI chatbots, including OpenAI’s ChatGPT and Google Gemini. Photo by Bob O’Connor
TerrierGPT Provides BU Community with Free Access to Leading Chatbots
New tool is the result of a partnership between the University’s Artificial Intelligence Development Accelerator and Information Services & Technology
As the world of artificial intelligence continues to expand, Boston University is offering its own chatbot for staff, faculty, and students: TerrierGPT.
The free generative artificial intelligence (AI) tool is the result of a partnership between BU’s Artificial Intelligence Development Accelerator (AIDA) for Academic and Administrative Excellence, an initiative tasked with exploring how AI technologies can be used in academic settings, and Information Services & Technology (IS&T).
TerrierGPT provides members of the BU community with free access to their choice of leading AI chatbots from OpenAI, Anthropic Claude, Amazon, Meta, and Google. The University’s community members can log into TerrierGPT at terriergpt.bu.edu using their Kerberos credentials.
Why bring chatbots to BU?
First, because AI’s impact on daily life is continuing to expand in profound ways.
“It’s become very obvious that generative AI is going to transform higher education and that the future workforce will need basic AI literacy skills on their résumés, independent of discipline,” says Kenneth Lutchen, BU’s vice president and associate provost for research and AIDA interim executive director.
“Our mission at BU is to create holistic citizens that get a degree in a specific discipline, but have foundational skills and capabilities,” Lutchen says. “We’re trying to ensure they know how to use generative AI in the most constructive, productive way possible for themselves, for their careers, and for society.”
It’s also a matter of equity.
“We saw an unevenness [across BU] with respect to knowing what AI [is capable of], and having access to AI models,” says John Byers, AIDA codirector and a Faculty of Computing & Data Sciences professor of computer science. “At a high level, the main goal of TerrierGPT is to democratize access to AI and give people access to a bunch of the best models out there.”
Finally, your personal and BU-related data are safer within TerrierGPT than outside of it.
“If you went to the free version of ChatGPT, for example, and entered your queries, that data has been sent to OpenAI and they can use it for whatever they want,” says Bob Graham, AIDA interim chief AI officer and IS&T associate vice president of enterprise architecture and applications. “For TerrierGPT, we established protections that mean any data entered into it is BU’s, and companies don’t have any right to that data.”
BU Today spoke to Lutchen, Byers, and Graham to get answers to some of the most commonly asked questions about TerrierGPT.
FAQ
What can TerrierGPT do?
TerrierGPT can be used for a variety of purposes, both personal and academic. For example, students can use TerrierGPT to generate study guides or model test questions, while faculty can use the tool to help create course syllabi or lesson plans. No one is required to use TerrierGPT, however.
Learn more about TerrierGPT use cases here.
Why offer access to different models?
Different models serve different needs. For instance, some models are better at logic and solving coding problems. Offering a variety of models allows users to select their preferred model or the model most appropriate for their tasks.
How should I approach using AI in an academic setting?
BU adopts a “critical embrace” theory toward generative AI in research and academia: that the technology should be utilized with sensible guardrails, while keeping in mind the benefits and limitations of AI. Overall, TerrierGPT should be used to augment, not replace, learning and instructional capabilities. Students and instructors should also be transparent about their use of AI.
Find AIDA’s generative AI guidelines for students and faculty and staff here.
Is TerrierGPT secure?
Yes. Unlike when you use non-BU versions of ChatGPT and other chatbots, the information you enter will not be used as training data.
AIDA’s website notes: “The platform complies with BU’s internal privacy and data protection policies—and none of the data entered is used to train external models. Data uploaded to the platform is only accessible by IS&T personnel and has the same strong privacy protections applicable to all BU enterprise data, such as emails and documents stored on BU’s OneDrive. However please note that TerrierGPT is not approved for use with restricted use data, including HIPAA-regulated information.”
What environmental concerns were factored into bringing this tool to BU?
Building an AI model from scratch requires a tremendous amount of energy. AIDA sought to leverage existing chatbot technology to significantly reduce the amount of resources needed to create TerrierGPT (something that BU is working on). The additional energy burden of using TerrierGPT is low.
What’s coming next for TerrierGPT and generative AI at BU?
There are future expansions planned for TerrierGPT, and related products, regarding new features and capabilities. BU also plans to launch an online generative AI literacy course for undergraduates, after which students will earn a digital certificate. For faculty, AIDA and the Institute for Excellence in Teaching & Learning are partnering to offer a series of symposiums this fall on using AI for instructional purposes. (Attendees must register for each event.)
Look for more updates from AIDA as the academic year progresses.
Find the answers to more FAQ about TerrierGPT, including information about technical specifications, here.