Dickinson Launches Lab Focused on AI Research
Where Curiosity About AI Leads to Questions Nobody Has Answered Yet
by Tony Moore
On June 23—the 114th birthday of computing pioneer Alan Turing—Dickinson will officially launch the Deep Network Understanding (DNU) Lab, a research initiative that gives students hands-on opportunities to investigate large language models and deep neural networks.
The DNU Lab is the brainchild of John MacCormick, professor of computer science and author of books such as Nine Algorithms That Changed the Future and the newly published Thinking AI. The lab's premise is straightforward: Students at any experience level can contribute meaningful research on some of the most consequential technology of the moment.
The Power Behind LLMs
"I'm fascinated by deep neural networks, which power large language models, AI assistants and most other modern AI tools," MacCormick says. "I wanted to create a community at Dickinson that could share in this fascination and investigate the myriad research questions that arise from deep networks."
The lab has been quietly building momentum since a soft launch earlier this year, which included a 16-student independent research course and eight student summer interns. Each semester, the DNU Lab will offer at least one research course in which students earn college credit—either half or full—while working on projects tied to a central theme. Prior experience with AI or neural networks isn't required; the lab provides onboarding support and pairs newer students with veterans.
MacCormick believes the lab fills a gap that extends beyond Dickinson's campus.
"I don't know of any other institutions that have a research lab specifically focused on providing undergraduate research opportunities in large language models and deep neural networks," he says. "The lab provides Dickinson students with a unique opportunity to position themselves well for the workplace or a research track in graduate school."
What AI Research Looks Like
That opportunity is already resonating. Aaron Shin '27 (computer science, mathematics) enrolled in the lab's inaugural research course last semester and returned this summer as an intern. His current project examines whether very small language models trained on short sequences can learn rules that generalize to much longer ones—a question about whether AI truly learns or merely memorizes.
"We are using simplified tasks and small models so that we can understand not just whether the model succeeds or fails but why," Shin says, noting that the project has allowed him to better understand what research in AI actually looks like day to day. "This work has given me a chance to connect coursework with real research, work closely with a professor and think more independently. I was a little surprised to see a lab like this launching at Dickinson, but in a good way. It is exciting to be part of it early on."
Into the Unknown
John Lee '27 (computer science), who joined after taking the same independent research course in his junior year, is investigating a related problem: whether small AI models can learn underlying rules rather than surface patterns, and which aspects of a model's design make that possible. The analogy he reaches for is intuitive—a student who memorizes solutions versus one who grasps the principle behind them.
Lee came to the lab with a background in hackathons and a growing interest in AI, and he's been thinking seriously about graduate school. The lab gave him a way to test that instinct.
"You don't see a lot of liberal-arts colleges standing up an AI research lab. I wanted to find out what research actually feels like, and this was exactly the chance to do that," he says. "In class I'm usually solving problems that already have a known answer. Here I'm working on a question nobody has the answer to yet, and I'm building things I had only read about in papers before."
That distinction matters: The DNU Lab isn't a simulation of research. And for students weighing graduate school, careers in tech or simply trying to understand what AI can and can't do, it's a rare chance to find out from the inside, from someone who literally wrote the book on the subject.
TAKE THE NEXT STEPS
Published June 22, 2026