In 2022, OpenAI released ChatGPT, the generative artificial intelligence platform that has fundamentally changed how we work, learn, create, read, and write. ChatGPT has become, to some extent, the default platform of reference in conversations about AI, especially when those who are conversing don’t know much about the technology. For example, students at the College are accustomed to hearing, “Don’t plug the prompt into ChatGPT” whenever a professor assigns an essay. As a humanities major, I hear it all the time.
The uncertainty that surrounds generative AI, large language models (LLMs), and similar technologies has given way to confusion in educational settings of every stripe—and a lot of that confusion has translated into avoidance. Don’t use ChatGPT. Don’t let AI do your work for you. Many of the humanities classes I’ve taken at the College have deliberately set up walls between students and AI, such that using an LLM in class quickly becomes academic dishonesty. But this practice runs the risk of insulating students from a technology that will play a massive role in their futures. It is not productive, I would suggest, to apply the humanities’ in-class mantra to your life. You should use ChatGPT and other LLMs, although maybe not on an English essay. You should become acquainted with generative AI.
Thankfully, the College has resources (lots of them) to make this happen. One of the most recent additions to its stable of generative AI resources is called MuleChat, and every member of the College’s community should know about it.
MuleChat is, essentially, a collection of generative AI tools available to students, faculty, and staff. ChatGPT may be the catalyst that set off the recent AI boom and a favorite LLM among students trying to shirk their homework responsibilities, but there are several other robust AI platforms on the market, like DeepSeek and Claude. The advanced models of ChatGPT and many of the other LLMs are also locked behind a paywall, which is often prohibitive; ChatGPT’s “Plus” plan costs $20 a month. MuleChat takes down that paywall and provides students with access to the best tools in the industry. Michael Donihue `79, who is the Herbert E. Wadsworth 1892 Professor of Economics and the interim director of the Davis Institute for Artificial Intelligence, described MuleChat as a way of providing “equal access for all people” to these industry-standard generative AI tools.
“We want everybody to be able to get access to the most advanced models,” Donihue explained. But MuleChat isn’t just about providing access. It is also a way for users, in Donihue’s words, “to learn the role of the human in the tool. When you [visit Davis AI], we don’t just turn you loose. We actually sit with you and help you craft a good prompt or see what a good prompt is.”
While writing this article, I attended a meeting of all the MuleChat tutors, who are the people with whom MuleChat users have to work in order to access the collection of generative AI tools. A few of them described the conversations they had had with MuleChat users. Many of those conversations were about the confusion that is prevalent in classrooms, syllabi, and in society more broadly—the confusion that has led to insular policies around AI use. What I like about Davis AI’s approach to artificial intelligence is its willingness to confront this confusion, rather than pushing it aside or jumping to conclusions. And by providing MuleChat to students, faculty, and staff, with the help of trained tutors, the College is providing us with the resources to address our own internal confusion in a productive way. I, for example, am not thrilled about the effect LLMs have had on writing-heavy industries, but at the same time, I recognize that my displeasure is best accompanied by a thorough understanding of the technology.
Davis AI’s mission lines up well with this overall goal.
“The mission [of Davis AI], technically, is to make Colby a leading institution where human-centered research and teaching of AI happens,” Donihue said. “So we want to be the leader in that space. And the human-centered focus aligns with the liberal arts.”
MuleChat is, obviously, a way of keeping up with the latest-and-greatest scientific developments. It’s a way of making sure that the College keeps its status as a with-the-times, tech-savvy, future-proof institution of higher learning. And those are valuable designations. But it’s also a way of making the platonic ideal of a liberal arts education reality. We students are supposed to be, according to the College’s mission statement, broadly acquainted with human knowledge, whether that knowledge takes the form of a book or a scientific development or a revolutionary technology, like AI.
To that end, MuleChat is more than just institutional lip service; it’s a way to acquaint ourselves with the knowledge of today. That’s why every member of the College’s community should know about MuleChat—and they should probably use it, or have a conversation with a MuleChat tutor. I may not be learning about generative artificial intelligence in my humanities classes, but I do intend to learn about it, and Davis AI provides the resources necessary to do that. To me, that’s also part of what it means to attend a liberal arts college.
Donihue spoke to the liberal arts mission of Davis AI and MuleChat, as well.
“The talks I’ve been giving lately sort of recall a time when some of the scientists of the 18th, 17th, and 16th centuries were poets, and they were philosophers,” he said. “They were called polymaths, because they did the scientific stuff, but they also had impacts in the arts. Then there was a break around the time of the Renaissance, when science went its own way and pursued rigor and paid less attention to the humanities. And the humanities went their own way, too. Right now, [generative] AI is calling us to come back together, and in the liberal arts, that’s what we do.”
~ Elias Kemp `27
Be First to Comment