Boston U. Denies AI Will Replace Teaching Assistants on Strike

Could artificial intelligence have prolonged the five-month-old strike that’s shut down much of Boston University? Or, more specifically, did a dean’s recommendation—an emailed suggestion that professors use artificial intelligence to manage all the course discussion and lab sections left without striking teaching assistants—galvanize the union’s opposition?

Representing roughly 3,500 graduate teaching and research assistants at the private university, the two-year-old BU Graduate Workers Union or BUGWU—also known as Local 509 of the Service Employees International Union—voted to authorize a strike on March 25, and its members still haven’t returned to classrooms.

In August, Inside Higher Ed reported that BUGWU’s strike had become the longest walkout of U.S. higher education student employees since at least 2014, eclipsing a record 147-day strike by University of Michigan graduate assistants during 2023.

Negotiations to improve pay and benefits such as health insurance and childcare had broken down after eight months of talks prior to BUGWU’s walkout. The union, which says graduate assistants won’t return to work without a contract, now hopes to sign its first collective bargaining agreement. After more than 30 meetings, the parties met for another bargaining session on August 7 to consider what the university said was its “final proposal.” No deal was reached.

It costs $90,207 for each of the 37,557 students to attend Boston University, and BU pays graduate student workers between $27,000 and $39,000 annually for “half-time” work up to 20 hours per week. But MIT’s Living Wage Calculator says that living in the Boston area requires a minimum of $62,000 each year. Investopedia says that in America as of July 2024, only five metro areas—greater New York City, Honolulu, San Jose, and San Francisco in the Bay Area, Orange County and Los Angeles in Southern California, and Washington, D.C—are more expensive places to live in than Boston.

However, the problem of low wages among teaching assistants isn’t limited to expensive metro regions. A disturbing 2020 report from the American Federation of Teachers shows that a quarter of the contingent and adjunct faculty at the nation’s colleges and universities rely on public assistance and 40 percent experience challenges with covering basic household expenses.

The dispute took a surprising turn when the dean of the College of Arts and Sciences, Dr. Stan Sclaroff, suggested that professors use artificial intelligence tools to help manage their course discussion sections and labs left without TAs because of the strike. A distinguished machine learning expert with a doctorate from MIT, Dr. Sclaroff taught for BU’s Department of Computer Science and then served as its faculty chair for five years starting in 2007. BU’s president then appointed him dean of the College of Arts and Sciences in 2019.

In an email to faculty, Dean Sclaroff recommended professors engage alternatives such as “generative Al tools to give feedback or facilitate ‘discussion’ on readings or assignments.” This suggestion quickly sparked outrage from the union, which shot back with a swiftly deleted post on X that said, “Yikes! Pretty telling of how little BU values our labor!” After critics accused BU of attempting to replace graduate assistants with artificial intelligence, the walkout rapidly drew coverage from the national media, and the resulting firestorm of controversy ensued.

BUGWU then responded to Dr. Sclaroff’s statement with a more tactful and diplomatic formal reply, saying that:

We are extremely disappointed by the university’s suggestion that the use of AI could even begin to substitute for the hard work that graduate workers pour into mentoring students, facilitating discussions, and teaching. We sincerely hope that the university would reconsider this suggestion and instead focus on properly compensating the people who do the work that is crucial in keeping the university running.

The administration’s damage control efforts quickly swung into operation. On March 28, the university issued a statement listing a broad range of teaching tools available to professors to “offset” a striking TA’s absence. These examples included combining discussion sections, offering discussion sections over Zoom video, and encouraging conversations through the discussion board function on the college’s Blackboard learning management system.

The university also said that “using technology, including generative AI, as part of discussion group activities is one of about a dozen possibilities that have been used in classroom settings.” The statement continued:

Neither Dean Sclaroff nor Boston University believe that AI can replace its graduate student teaching assistants, and the assertion that we plan to do so is patently false. Everyday hundreds of discussion groups are being led by teaching assistants and that is not going to change.

The administration also said it looked forward to “continuing to negotiate with the union and to the return of those teaching assistants who walked off the job when the strike began.”

But the public relations damage had already been done. A member of the faculty who spoke on conditions of anonymity with the New York City-based Daily Beast—the outlet that broke this story—remarked that “for some bewildering reason, they decided to throw in an extremely non-conventional and ultimately self-damaging suggestion that we just use ChatGPT to do the work. It’s honestly pretty shocking.”

“It’s demoralizing, for sure,” the faculty member added. “You have critical, interesting, and accomplished professors who are, like me, baffled by this belief that you can just deliver a quality higher education experience through the punch of a button.”

Overall, professors were none too pleased that Dean Sclaroff had suggested that they should rely on AI for support with their absent TAs’ discussion and lab sections. For example, Dr. Nathan Phillips, a physiological ecologist in the Earth and Environment Department, told Boston’s Daily Free Press that professors will not follow his recommendation because they don’t believe artificial intelligence provides an “authentic education:”

Education is more than just the transmittal of information. It’s about being in a community of learners and teachers. That suggestion is the antithesis of everything that the concept of education stands for.

Across town at Northeastern University, Dr. David Aldrich, a professor of political science and public policy, took to X and fired off “Robot scabs? Boston University administrators recommended that faculty members use generative AI tools in classrooms due to the BU Graduate Workers Union strike.”

Undergraduate students also expressed alarm over the Dean’s recommendation for professors to adopt AI. College of Arts and Sciences freshman Ellie Winkler Quigley lamented to the Free Press that she wouldn’t be able to develop personal relationships with teaching assistants if her classes switched to AI tools.

“With the TA, you’re able to connect with them,” she said. “With AI, that completely takes away the ability to connect with people. That just puts everything on a computerized level.”

“There’s a human element of having your grad worker give you advice on a piece of work that AI can’t do, and that is just essential for learning to be productive,” said Vijay Fisch, a sophomore in the Center for Computing and Data Science. Because AI won’t reach the high level of writing of BU graduate workers, Fisch called the recommendation to use AI in the classroom “absurd.”

Because of the pro-labor stance of the Biden Administration—along with other factors such as pandemic health concerns, labor shortages, and a more favorable view of unions by younger workers—U.S. colleges and universities have experienced a sharp spike in labor actions since 2020. These include strikes among graduate teaching and research assistants like those at BU and Michigan, along with other forms of labor unrest as well.

“We are seeing a ferment of strikes and strike threats in the United States by workers off and on campuses,” said labor expert William Herbert. He’s a public policy professor and the executive director of the National Center for the Study of Collective Bargaining in Higher Education and the Professions at the City University of New York’s Hunter College.

Because of changes in its appointed membership, the federal agency that enforces labor laws in the United States—the National Labor Relations Board—sharply reversed its policies on graduate student unionization efforts in 2017 and 2021. The NLRB encouraged graduate students to unionize during the Obama and Biden administrations but refused to allow unionization by grad students during the Trump Administration.

Despite all the controversy and the fierce criticisms of Dean Sclaroff’s remarks, it’s important to recognize that artificial intelligence continues to make inroads at augmenting the functions of educators across the United States, especially in elementary and secondary education. For example, our feature article ”Are California Teachers Actually Using AI Software to Grade Papers?“ recently covered the new trend where California teachers are using AI platforms to grade essays and give students feedback on their writing assignments.

The trend appears to be catching fire thanks to billions of dollars of AI investments from the tech and venture capital industries in Silicon Valley, which have targeted education as the first “killer app” for artificial intelligence.

Meanwhile, the Boston University strike’s national press coverage likely led to legislative initiatives in at least two states intended to prevent artificial intelligence from taking the jobs of college faculty. For example, as we point out in our article California Bars AI From Replacing College Instructors, California’s Governor Gavin Newsom signed into law AB 2370 in July 2024. This landmark legislation ensures that only qualified human instructors teach courses in the California Community College system—thus preventing generative artificial intelligence platforms like ChatGPT from functioning as instructors.

In Minnesota, a similar bill would prohibit universities and colleges in the state from using a generative artificial intelligence platform as “the primary instructor for a credit-bearing course.” Although the proposed legislation has a broader scope than California’s AB 2370 because it would regulate not only community colleges but any college or university in the state, the Minnesota bill’s language doesn’t appear to block colleges from using AI as secondary instructors. The absence of such a provision would enable Minnesota’s universities to replace graduate teaching assistants who teach discussion sections and grade assessments—just like the 3,500 TAs still on strike at Boston University.

The threat that AI technology could take teachers’ jobs also concerns K-12 educators. On July 5, the largest teachers union in the United States—the National Education Association with 1.4 million members—voted to approve an educational policy statement that advocates restrictions on the use of artificial intelligence in elementary, middle and high schools. Human teachers should “remain at the center of education,” says the statement.