Refusing AI in Higher Education

Smart University: Student Surveillance in the Digital Age by Lindsay Weinberg. Johns Hopkins University Press, 2024.

A few weeks ago, I received an email from De Gruyter Brill, the publisher of my first book, informing me that the text, along with my other journal articles and book chapters that Brill has published, will be used to train AI large language models. The company offered me neither compensation nor a way to opt out. However, according to Lindsay Weinberg’s new book Smart University: Student Surveillance in the Digital Age, academic workers do have a choice. AI might have arrived in higher education without invitation—but not without resistance.

If the ed-tech industry is to be believed, AI will entirely transform higher education and improve “student success” by making the sector more customizable, equitable, personalized, and efficient. There is a fundamental threat behind all this PR. With the arrival of Google’s new NotebookLM—which can generate notes and summaries of any book or document and convert text into audio to create podcasts—the industry stands ready to use AI to make the teaching and research of professors obsolete.

But it won’t be smooth sailing for the industry if Weinberg, clinical assistant professor and director of the Tech Justice Lab at Purdue University, has anything to say about it. Weinberg foresees the potential for students and faculty members to develop “solidarity around resisting smart university initiatives.” The insertion of AI into higher education is only the latest effort of corporations to force technology onto our campuses, and students and faculty have long resisted and disrupted such attempts. The Berkeley Free Speech Movement, for example, issued a clarion call for students to refuse to use computer punch cards, an increasingly common method of data storage in the 1960s. Student activist Mario Savio anticipated concerns about commodification of data when he said during a 1964 sit-in at the University of California, Berkeley, that students “don’t mean to end up being bought by some clients of the university.”

The strength of Smart University is Weinberg’s incisive critique of AI’s core myths of customizability, personalization, equity, and efficiency as actually rooted in efforts to standardize education. Rather than customizing and personalizing higher education to improve learning and equitable access, AI is a racialized and gendered technology for carrying out mass surveillance of students “from cradle to career.” Weinberg shows how the data swallowed up by AI, as with previous forms of AI and algorithmic management, are extracted from the unpaid labor of students, faculty, and other users. These data are used to control, discipline, and manage us in the university and beyond. And, in the process, the data feeding the “smart university” are produced by and reproduce the existing systems of exploitation and domination based on class, race, and gender.

The book demonstrates how AI’s widening use for student recruitment (the focus of chapter 2), student retention (chapter 3), wellness monitoring (chapter 4), and campus security and exam proctoring (chapters 1 and 5)—as well as in academic research—is hardly an aberration. Weinberg authoritatively recounts the critical history of higher education as rooted not only in subjugation of the poor, the working class, and people of color in the United States but also in efforts to acquire knowledge for global colonialist domination. After student movements ratcheted open and diversified universities in the 1960s while confronting institutions’ roles in war and imperialism, neoliberal institutional restructuring manufactured a fiscal crisis as states disinvested from higher education. This merging of ed-tech with the university is a “stitching together of capital, the state and academy in ways that further militarism and mass incarceration.”

Weinberg shows how ed-tech companies market AI by exploiting the austerity, downsizing, and outsourcing of higher education precipitated by neoliberal restructuring. “Datafication” of education has become the strategy for extracting ever more labor from faculty and students to increase measurable productive outcomes such as grade point averages and graduation rates. Smart University methodically deconstructs the marketing language by which ed-tech sells itself to a ballooning layer of campus management: for example, by “customizing” programs (which really means bypassing shared governance bodies and unions by leaving them out of the decision-making process); “personalizing” education (or alienating students from one another); increasing “efficiency” (or increasing the workloads of the precarious faculty supermajority); and improving “student success” (or accelerating students’ progress through programs at ever greater speeds).

The obsession with “campus security” evokes the relentless concern of neoliberal management with securing the college or university from the uncertainties created by neoliberal restructuring. The supposedly “seamless” generation of real-time data on every action of students and faculty through ubiquitous mass surveillance is built to coercively extract labor and speed up graduation. Weinberg documents the direct line from the development of facial-recognition AI to AI-driven proctoring software as a means of policing and controlling student behavior and discusses the racial impacts of doing so. Not accidentally, the “smart university” is modeled after the “smart cities” model that approaches development, growth, and predictive policing with the same reliance on data generated by massive “dataveillance” to control, discipline, and manage racially diverse working-class populations and prevent political contestations and ruptures.

Those concerned with the disparate impacts of AI have done little besides demanding privacy and transparency to contain and manage its threats. For Weinberg, this liberal framing of AI in terms of ethical and privacy issues is inadequate and ineffective. Since a 2008 amendment to the Family Educational Rights and Privacy Act, universities have been allowed to “share” student data with corporations. Weinberg proposes that, rather than trying to opt out or regulate access to these data, “the most ethical choice is refusal.” We must begin organizing against AI as an issue of control over our work and workplaces.

Weinberg’s incisive critique of the forces seeking to automate higher education leaves me with questions for further examination. This short book trains its sights on the restructuring of higher education “around production and capture of data.” What exactly are data being captured for? Is it simply to surveil for the purposes of tracking and coercing behavior, driving consumption, and managing racially diverse populations? Or is it an emerging labor management strategy that socializes and trains workers to work obediently as what Karl Marx called an “appendage” of the machine? Are the products of the automation of higher education merely data, or are the data being generated to produce more disciplined and productive workers?

By the end of Smart University, I wanted to know for what exactly the university is being restructured. If, as Weinberg persuasively argues, the university has always been an integrated institutional partner in extending capitalist domination at home and globally, then this materialist reading can also be applied to the intended outcomes for AI-infused institutions. Is the “smart university” only the latest strategy in harnessing higher education to capital, or is it something fundamentally different? Weinberg connects AI to “big data socialization” that aims to produce workers who better serve capital’s needs for educated skilled labor, but this is surely only one of many outcomes. Smart University’s argument would be strengthened by emphasizing how the de facto objective of the smart university, along with all other management strategies and technologies forced on higher education, is to produce ever more disciplined, productive, and obedient labor power.

Smart University has taken us to the edge of the precipice, asking us to look into the abyss that awaits us if we do not organize and escalate our refusal of AI. To inspire the effort, Weinberg reminds the reader of the 2019–20 University of California, Santa Cruz, wildcat strike in which graduate students calling for a cost-of-living adjustment refused to enter grades in the online Canvas learning management system. Calling attention to AI’s role as part of a system of labor control can help us begin to strategize the forms that further refusal will take. Thoroughly researched and expansive in its scope, Smart University is an indispensable critique of the forces seeking to automate and dismantle higher education as we know it. There is simply no other book like it. We ignore Weinberg at our own peril.

Robert Ovetz, senior lecturer in political science, teaches labor relations in the Master of Public Administration program at San José State University and is author and editor of five books on the labor movement. His email address is [email protected].