Ideas Made to Matter
Artificial Intelligence
AI Expert Spotlight: Vivek Farias
By
We asked several MIT experts about their latest projects and what they see as the most exciting—and concerning—aspects of the AI boom.
Vivek Farias, Professor of Operations Management
is using massive parallelism—many processors working simultaneously—to speed up different types of computations. He is also studying the intersection of large language models and human behavior to simulate consumer behavior and determine the impact of bias on LLMs.
How would you describe your work in the artificial intelligence space?
I’m working on a few things. One is using the massive parallelism implicit in the core computational machinery that enables deep learning — for instance, to simulate all of Walmart's supply chain operations in minutes as opposed to months. Another area I am deeply interested in is the intersection of large language models (LLMs) and human behavior. LLMs can simulate consumer behavior, which is quite valuable — imagine all your customers in a box. At the same time, LLMs themselves are trained with certain implicit models of human preference, and I am studying the impact of those models on what the LLMs themselves might do—for instance, in their implicit biases.
What are you most excited about in working with AI?
LLMs have the ability to fundamentally change how we live, work, learn—you name it. This is such a momentous change that I am happy to play any role I can in realizing this vision. Scientifically, I am probably most interested in the intersection with human behavior.
What do you see as the biggest challenges or areas for caution?
Probably the implicit biases we’re building into LLMs.
READ MORE IN OUR AI EXPERT SPOTLIGHT SERIES
