Nathaniel Rodwell-Simon

Edward H. Levi Hall, which houses many of the University’s administrative offices.

The Council of the University Senate will meet on Friday to continue discussions over the creation of a new Division of Computational and Mathematical Sciences.

The new division, whose establishment was recommended by a faculty committee in an October report, would initially include the Departments of Computer Science, Mathematics, and Statistics. It could eventually house a new AI institute, offer new degree programs, and require additions to the College’s Core requirements.

Discussions about creating a new division have been ongoing since at least May 2024, when a previous report by the Committee on Advancing Computing and AI recommended that departments “central to AI [and machine learning] foundations” split off to form a “new decanal unit.” The October report outlined the importance of a new division that could better support other disciplines across the University as AI becomes increasingly foundational to academic research and would “affirm the University’s commitment to intellectual leadership in a world increasingly shaped by these fields.”

The new division “would be good for the University as a whole in the sense that [its] fields are increasingly important across a range of human intellectual activity right now,” Hank Hoffmann, chair of the computer science department, told the Maroon.

To pass, the proposal would need a majority vote from the 51 elected members of the Council of the University Senate. That vote, however, would be subject to veto by the University president and the final decision would rest with the Board of Trustees.

If approved, the new division would be a significant change to the structure of the University, which has been composed of four divisions, Biological Sciences, Social Sciences, Physical Sciences, and Humanities—renamed Arts & Humanities last year—since 1930. The Physical Sciences Division currently houses the computer science, mathematics, and statistics departments. A new division would also create “an opportunity to explore ways to further integrate computational thinking into the mathematical sciences sequence in the Core,” according to the October report.

Standard operating procedure in the Council is to hold two meetings on a potential decision and take a vote at the second meeting, according to David Miller, who serves on the Committee of the Council, a managerial body of the Council. Although the Council completed a first reading of the report on December 9, making the coming Friday the second reading, Hoffmann said he was unaware of any plans for a vote.

Some faculty members have expressed concern about the quick timeline and the lack of revisions incorporated into the report recommendations despite feedback the faculty members offered.

Miller, who also participated in an April 14 “panel of supporters and critics” of the proposal, shared a number of concerns with the Maroon about “the lack of robust discussion amongst the faculty and a simultaneous lack of serious assessment of the broad implications that creating such a new division might have.”

This Friday’s meeting was not initially scheduled, according to Miller, and is being held because the Committee of the Council insisted on it. The committee was “shocked that [a meeting] had not been built into the schedule as per standard operating procedure,” he said.

Miller also noted that there have been no revisions to the report despite what he described as “significant feedback” from faculty. Though it is “normal matter of course” for reports to be revised, it is not a formal requirement, and a vote could proceed without it at either of the upcoming Senate meetings, Miller said.

“It’s been ~30 weeks since the report was finalized and there have been 5 Council meetings since the report was delivered to the Council, and not so much as a footnote has been added,” Miller wrote in an email to the Maroon.

According to both Miller and the October report, a primary concern for many faculty members is a financial one—both the potential for weaker fundraising in a smaller Physical Sciences Division and the significant cost of launching a new division amid significant budget issues.

“There are people in the Physical Sciences Division who are worried that this will dramatically financially injure the remaining departments in the Physical Sciences Division,” Miller said. The report mentions the need for “careful institutional understanding and attention” to budgetary issues but does not address how they would be resolved.

Miller also noted that although the May 2024 report listed “a more comprehensive study exploring intellectual and financial tradeoffs” of creating a new division as its first “major” recommendation, no such study has been conducted. Financial concerns “[weren’t] something that the committee really investigated,” according to Hoffmann.

The Physical Sciences Division receives a significant share of its revenue from the computer science and data science departments, which have the division’s largest undergraduate enrollments and operate its largest master’s programs. Computer science, data science, statistics, and computational and applied mathematics master’s programs account for 79 percent of master’s enrollment in the Physical Sciences Division, according to University enrollment reports. “Financially, [computer science] and data science…are generating well over $100 million per year in tuition, research awards and gifts,” computer science, data science, and statistics department chairs and faculty wrote in an op-ed in the Maroon in September.

The Massachusetts Institute of Technology; the University of California, Berkeley; and the University of Wisconsin–Madison have all announced new academic divisions in computer science in the past decade. At the April 14 panel, computer science professor Ben Zhao commented on lucrative job offers he had received from other universities, according to multiple faculty members who were present. It is common in computer science for top faculty to be poached by universities in a bid to attract talent.

There has also been a broader explosion in salaries for academic researchers in technology. According to a March report released by the Becker Friedman Institute, more than two-thirds of AI researchers work in industry, up from less than half in 2001. While top industry earnings nearly quadrupled between 2001 and 2021 to $1.94 million in 2015 dollars, both top and average academic salaries have barely changed.

In the October report, Hoffmann and his colleagues wrote that “[i]naction would be an existential crisis for the University, not just in core AI, but across nearly all disciplines of academic inquiry,” comparing Al’s integration into basic research to the introduction of microscopes in the natural sciences.

Regardless of whether a vote happens in the next month, Hoffmann said there is “a hunger” for collaboration with researchers across the University. “I’d like a new division like this to have a deputy dean whose job it is to try to facilitate these collaborations.”

“My colleagues in other units of the University want to work with computer scientists. We have computer scientists that want to work with the rest of the University,” Hoffmann said.