Matthew Jones Studies the History of Data, Surveillance, and AI

As a kid, Matthew Jones could often be found sitting in his family’s kitchen, programming on one of the first personal computers. At the same time, the native of Reno, Nevada, was deeply interested in the humanities, particularly history and public policy.

It wasn’t until he took a course on the history of science his freshman year at Harvard that Jones realized he could blend his passions. “The combination of my interests was to think about math historically and philosophically,” he says.

Jones went on to earn his Ph.D. in history and science from Harvard, as well as his M.Phil. from Cambridge. He focuses his work on the history of information technologies and intelligence, early modern science, and data collection.

“We as a society build intellectual and technological systems that come to take on such an imposing presence that we think things can’t be otherwise. It’s something we need to be able to analyze critically,” Jones says. “The point is not to call into question everything around us. Rather, it’s to draw upon historical ways of thinking to understand how things came about and to understand that there is the possibility they could be otherwise.”

Quick Facts

Title
Smith Family Professor of History

Time at Princeton
2 years

Recent Class
The Scientific Revolution

Jones’ Research: A Sampling

Image

Mikel Casal

The Human Element in Data

For seven years at Columbia University, Jones co-taught, with Chris Wiggins *98, a course on the history of quantitative reasoning and machine learning from the 1800s through the modern AI era. In 2023, he and Wiggins published How Data Happened: A History from the Age of Reason to the Age of Algorithms, named a New Yorker Best Book, which expands on their course and probes the social and political contexts that produced and fundamentally shaped these tools. “The development of statistics was deeply enmeshed in the social problems and ways of thinking of its day. Biases are baked into the datasets, and those tend to be replicated by predictive algorithms,” Jones explains. “How then do we deal with these algorithms and models as they become evermore central to government, corporate, and academic institutions?”

Surveillance State

Starting in the 1980s and expanding after 9/11, the United States has engaged in state-sanctioned digital monitoring of residents’ communications. In his next few book projects, Jones will explore the evolution of this surveillance and the gulf between technological advancements and the law — particularly around privacy and the Fourth Amendment. “How was it that in the immediate wake of 9/11, the U.S. government stood up a massive surveillance operation that included U.S. persons? How was that possible politically, technologically, legally?” he asks. While the prevailing argument is that technological advances necessitate specific legal changes to balance privacy with security, Jones argues that “the internet and generative AI don’t produce that law. That distinction is one of the key things I’m interested in.”

How AI Fits Into Our World

This spring, for the second time, Jones is co-teaching the course Data and Culture with English professor Meredith Martin. By first examining the history of data and analytics from both a quantitative and humanistic perspective, the course questions what the growing centrality of large language models (LLMs) and AI will mean for our world. “How do we think about the coming of machine learning and generative AI as a question of labor, a question of knowledge, a question of practice?” he asks. “We’re really aiming to provide students with an armature of ways to understand not just the technology, but also how it fits into the knowledge and work landscape.”