When capital recombines, workers’ inquiry is a tool for understanding its new form and the resulting class composition. First articulated as a never-used quantitative survey by Karl Marx, workers’ inquiry became legendary when the Italian operaisti used the method for analyzing the 1970s Fiat factories.^[1] Now, in another moment of recomposition, there has been renewed interest in the method. Publications like Viewpoint Magazine, Notes from Below, and, more recently, Longhaul have spread theory and application in workers’ inquiry. After some halting starts—like Voices from the Valley, a 2020 product of Logic magazine’s middling collaboration with Farrar, Strauss, and Giroux—academics studying the tech industry have also rediscovered the approach. For example, the Distributed AI Research Institute (DAIR) cites Notes from Below editor Jamie Woodcock in their outline of a new data workers’ inquiry project. From a similar milieu, the Capacitor Collective published Notes Towards a Digital Workers’ Inquiry to further solidify a project of “collaborative research that feeds into labor organizing within and against digital capitalism” (pp. 1). The result is a productive, if limited, compilation.

A guiding essay provides the theoretical framing for the eleven following interviews about the tech industry, research labor, and related organizing efforts, like the Tech Workers’ Coalition, The Workers’ Observatory, and Turkopticon. In this essay, the Capacitor Collective explain that they are working against a trend of academic researchers who “parachute into a labor community […] in order to extract knowledge from workers” (pp. 10). The collective emphasizes that digital workers’ inquiry is “not a research method but rather a political approach to the production of knowledge” (pp. 32). Knowledge production is a load-bearing element of tech organizing campaigns, because “an inability to understand the ‘calculative logics’ that underpin algorithmic systems is a defining feature of the conditions faced by the digital workforce and of its technical composition” (pp. 21). Understanding, for example, how rideshare apps assign jobs can help drivers identify forms of concerted action that will improve their conditions, e.g., all logging out of the app at the same time to drive up algorithmically-offered rates within a given area.

I appreciate the immense work that Notes Towards comprises and the importance of what it represents—a veritable attempt by academics to develop and demonstrate a relationship and shared project with social and labor movements. The choice of presentation and method, however, has me worried. Marxist academics have spent a decade or so revitalizing workers’ inquiry as a bottom-up process of knowledge production, useful to workers themselves. Where it intersects with labor tech academics, however, I fear this effort will be subsumed by the immense engine of knowledge, and therefore consensus, production that tech capital has developed. Notes Towards reflects a subtle redefinition of workers’ inquiry that should be resisted and that is, in fact, resisted by the data the book contains: we should resist any appropriation of workers’ inquiry that severs it from its theoretical roots and practical applications. I am highlighting those elements, so that the next attempt can go further, farther. This requires a renewed interest in the roots of workers’ inquiry. I will start with the problem of how writing comes from and represents workers.

Workers and Their Written Representatives

This is not an uncommon dilemma in the tradition of workers’ inquiry, one that relates to questions of method and representation. The Johnson–Forrest tendency ran into this problem with Phil Singer’s 1947 pamphlet The American Worker. Singer wrote about his experience as an auto worker in a New Jersey GM factory and uses that experience to argue for the wildcat strike as a necessary venue for the development of workers’ autonomy at the time. As Haider and Mohandesi point out, however, “The openness of [Singer’s] narrative form exaggerates a tendency to slip from measured generalization to untenable overgeneralization. By trying to fuse his subjectivity with that of the rank and file as a whole, Singer ends up attempting to legitimize himself as a reliable mouthpiece for all the workers in his factory.”^[2] Marx’s original proposal was for a structured survey, but worker inquiries now tend to take a narrative and ethnographic form, which provide rich subjective accounts but make for difficult work in broader analysis. This “openness” is necessary because it continually refocuses workers’ inquiry on the subjectivity of workers, rather than merely producing data about the labor process that can be expropriated for managerial control.

But there are many possible solutions to the methodological problems that open-ended inquiry brings. Notes from Below, for example, launched its Class Composition project in 2021 to understand economic changes wrought by the Covid-19 pandemic; they focused on collective written contributions and a more sectoral data collection and analysis. As Notes from Below editors explained, this research design reflects their prior difficulties both with locating “an identifiable militant layer that sees itself as such and thinks politically about its workplace” in certain sectors and then responding to the resulting unevenness in contributions to their enormous project—analyzing the entirety of class composition in the UK.^[3] Notes Towards a Digital Workers’ Inquiry is concerned with a single category of work (i.e., the digital) but faces a similar problem in the scope of that category’s inclusion: warehouse workers, gig drivers, precarious academic researchers, programming engineers situated across continents.

The mind boggles at how all these types of worker can be considered together, but that is the work that faces us. Notes Towards does not address, however, its resulting problems with method since its primary concern seems to be proposing workers’ inquiry as a method at all. Before I continue this line of thought, I have to describe the results of its inquiry, such as they are. Some of the work I’m doing here is necessary because, outside the framing essay, there is no analytic work done to draw together (or differentiate) the interviews.

Conditions of Digital Work and Its Inquiry

Notes Towards does provide numerous examples of how workers’ inquiry benefits strategic research, compared to traditional social science. Organizers with the Turkopticon project first surveyed academic research on the shared conditions of Amazon Mechanical Turk workers, only to realize that a browser plugin being used to track wages was deflating estimates of their hourly pay. But wages were not the central issue for Turkers: “There seems to be a narrative researchers use that characterizes Turkers as poor, impoverished, and unskilled workers who make $2–$3 per hour. […] We found that to be troublesome because we don’t see ourselves doing slave labor or busting our rears for $2 per hour. Nor do we see ourselves as being unskilled” (pp. 64). When workers conducted their own survey, they identified that “mass rejections” of completed tasks—allowing contracting companies to deny pay—were an issue of higher concern for Turkers.

Through this realization, they also changed their understanding of their shared subjectivity—organizing became a fight for recognition and dignity, not primarily against immiseration. A researcher working with gig workers in Edinburgh similarly observes that for many migrant workers, these positions represent significant increases in their hourly wages. While flawed academic research still circulates with narratives about “immiserated workers,” Turkers have worked to change that perception among themselves. In another interview, an Amazon worker–organizer describes how assumptions about the infamous fragility of last-mile facilities in the supply chain mistakenly predispose organizers to focus there, instead of on the more populous “first-pile” centers, where there are more workers massed.

But the longer you look at Notes Towards, the more the volume seems concerned with the conditions that produce research about digital work, rather than digital work itself. “Notes Toward” is an accurate qualifier for the Capacitor Collective’s project, but it also raises a question. Is Notes Towards an account of inquiries into digital gig-economy and platform labor, or the production of labor tech research itself? On this question, the book seems divided. Mechanically, out of the 27 total contributors listed at the back of the book, 18 are academics, researchers, or graduate students. (Little information is included about how this sample was selected.) When you count out the handful of others who present niche perspectives on inquiry—a union staffer, a union president, the founder of a workers’ center—we are left with very few direct accounts from digital workers and far more discussion of how academic research is conducted. A few interviews stand as examples of what components of fuller inquiry would look like, while others wrestle with standard academic ethics-wrangling about power imbalances between researchers and workers.

It is possible, however, to discern a clear proposal for how digital workers’ inquiry should be done that is present across some of the more insightful interviews: the division between researchers and workers must be broken down: “There’s a view that workers are not intellectual. Like the scholars are intellectual but the workers are not. That’s a false divide” (pp. 44). As one Tech Worker Coalition respondent pointed out, writers were most useful in conversations directed at inquiry when they were present as workers with their own workplace struggles. In one tech workers’ circle he led, a Buzzfeed journalist joined not as an outside ear but to bring her own experience of work to the discussion. The challenge then lay in connecting her experience to the wider circle of tech workers, who are not under the exact same pressures as journalists.

Conditions of Knowledge Production of the Same

Research methods can be tailored to realize new issues for a campaign to organize around (more than just providing organizers with an “ammunition” of stock phrases as one interviewed staffer says). For example, an Amazon worker explains how a map of bodily injuries highlights how particular pains are actually collective: “People put dots on the body map and you can see that everyone’s upper backs hurt. That quickly becomes a way to talk about pain and work where you can see that everyone’s upper backs hurt” (pp. 133). Alternative models are also proposed in Notes Toward when it is not possible to remove divisions between workers and researchers. The Workers’ Observatory in Edinburgh provides research support to local gig workers, rooted in the labor history of the city. As the founder describes, researchers can direct funding towards workers in ways that develop their collectivity and autonomy.

That is, in fact, the entire point of workers’ inquiry and its often-synonym, co-research. In an early formulation of the project that would lead to Notes from Below, Jamie Woodcock states,

The aim is to move towards an inquiry ‘from below,’ a form of co-research that breaks down the separation between researcher and subject. At its core the project is one of knowledge production and political organisation, and there has to be an awareness of this tension. The workers’ inquiry cannot simply be limited to an academic tool for refreshing theory.^[4]

To distinguish itself from management research that extracts data from workers and reinforces capital’s control over labor, workers’ inquiry (especially when supported by academic researchers) needs a radical commitment to its political project. Although they acknowledge that necessity, I am not sure that is what the Capacitor Collective are doing.

For example, one academic influence that reappears in the collection is the Distributed AI Research Institute (DAIR), which Timnit Gebru founded after she was fired for publishing an academic critique of large-language models. DAIR has become a home for several former Google ethics researchers, as the company clamps down on its proffer of academic freedoms. Looking at their recent collaboration on a Data Workers’ Inquiry, DAIR and its research collaborators elaborate a Workers’ Inquiry as a Research Methodology (WIRM). I am skeptical. WIRM’s authors do not engage deeply with Marxist theory to explain their use of this approach. In their definition of WIRM, for example, they seem to ascribe operaismo—the Italian Marxist tradition from the 1960s—with proposing the development of participatory action research, an academic methodology with its own connection to Marxist theory via Paolo Freire.^[5] That proposal was made by Woodcock in his 2014 article, not by Mario Tronti, Romano Alquati, or Raniero Panzieri.

The DAIR researchers’ described limitations for the approach also emphasize the distance between researcher and worker roles and goals, with “the sharp power differentials between ourselves, as relatively privileged, global-North-based researchers, and the data workers [they] collaborated with.” Rather than breaking down divisions between types of labor, this reconceptualization of workers’ inquiry reinforces those divisions—divisions that appear as insurmountable gaps in perspective and purpose. Notes Towards takes a similar, although not identical, approach to the adoption of workers’ inquiry. The Collective acknowledges the cataclysmic pressure that the tech and philanthropic capital applies to knowledge production about the tech industry: “Companies pour millions into think tanks that heavily influence the conversation around the role of digital technology in our societies. They hire researchers, seed research institutes and internal AI ethics teams, and fund conferences and research chairs at public universities” (pp. 11). In a footnote, they single out Microsoft Research for their role in that project but do not mention that one of the central interviewers is employed by the Data & Society think tank, which was founded as a project of Microsoft Research and has continued to receive continued foundation support from similar sources.^[6]

I mention this not to condemn the project of Notes Towards a Digital Workers’ Inquiry but rather to point out that this is an unresolved contradiction within it. Workers’ inquiry is for politics, not ethics. Academics are wondrously skilled at the worried ethical stance, the consummate professional who knows all the nuanced details wrong with what they do and does it anyway. Entire disciplines have developed on academics’ continually reciprocal reassurance that the unethical is ethical and, what’s more, necessary.

Over the last decade and with the assistance of hungry moral philosophers, the Association of Computing Machinery’s Fairness, Accountability, and Transparency (FAccT) conference has created intricate ethicswashing frameworks that discursively contain and neutralize the many violences done by machine learning and large-language model algorithms.^[7] As the name of the conference suggests, these frameworks have most often relied on virtue ethics and then try to develop quantitative measures of those values that can be operationalized in code or engineering as, e.g., “values in design.” The distance between the highly abstracted social values and how they are implemented or measured is huge. And, as political scientist Ben Wagner argues, humanistic ethics frameworks like this actually form flimsy substitutes to industrial regulation: “The idea that ethical frameworks provide a way to go beyond existing legal frameworks is taken to mean that in some cases this may also mean ignoring them.”^[8] FAccT is also financially subsidized by the largest tech corporations, so that products like algorithmic hiring tools can be anointed by academics as without bias or discrimination.^[9] For a time, FAccT was a site of academic struggle, but the result is research as reformism. I do not believe FAccT’s founders set out with that intention, but the pressure that research funders exert can bend any path, in time.

To forestall such a fate, Marxist researchers need to keep their purpose and relations constantly in mind. A framework like WIRM, I believe, too easily rips workers’ inquiry away from its radical roots. Workers’ inquiry is not a way to balance power between researchers and workers. It is a means to put the production of knowledge into the hands of workers, so they develop themselves collectively and their own solutions to the problems they face. Notes Towards is that, in medias res, for a particular class fraction of academics. here

1. Asad Haider and Salar Mohandesi, “Workers’ Inquiry: A Genealogy,” Viewpoint Magazine, September 27, 2013, https://viewpointmag.com/2013/09/27/workers-inquiry-a-genealogy/.

   ↩

2. Asad Haider and Salar Mohandesi, “Workers’ Inquiry: A Genealogy.”

   ↩

3. Sam DiBella, “Interview with Jamie Woodcock and Sai Englert of Notes from Below,” Washington Socialist, September 2021, https://washingtonsocialist.mdcdsa.org/ws-articles/21-09-notes-from-below.

   ↩

4. Woodcock, Jamie. 2014. “The Workers’ Inquiry from Trotskyism to Operaismo: A Political Methodology For Investigating the Workplace.” Ephemera 14 (3): 493–513.

   ↩

5. Goñi, Iñaki. 2025. “Communicating Knowledge Otherwise: Reclaiming Latin American Participatory Action Research to a Global Citizen Assembly.” Science for the People Magazine, November 5. https://magazine.scienceforthepeople.org/vol27-1-rethinking-scicomm/communicating-knowledge-otherwise/.

   ↩

6. Data & Society. “Funder List.” Accessed December 29, 2025. https://datasociety.net/funder-list/.

   ↩

7. To see an example of the kind of contortions made to justify disciplinary collaborations between computer science and moral philosophy, see: Elettra Bietti. 2020. “From Ethics Washing to Ethics Bashing: A View on Tech Ethics from within Moral Philosophy.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, January 27. World. https://doi.org/10.1145/3351095.3372860. For a description of how this refined process has already captured the “AI safety” community, see: Ahmed, Shazeda, Klaudia Jaźwińska, Archana Ahlawat, Amy Winecoff, and Mona Wang, “Field-Building and the Epistemic Culture of AI Safety,” First Monday, April 2024, https://doi.org/10.5210/fm.v29i4.13626.

   ↩

8. Wagner, Ben. 2018. “Ethics as an Escape from Regulation: From Ethics-Washing to Ethics-Shopping?” In Being Profiling. Cogitas Ergo Sum, edited by Mireille Hildebrandt. Amsterdam University Press.

   ↩

9. Young, Meg, Michael Katell, and P.M. Krafft. 2022. “Confronting Power and Corporate Capture at the FAccT Conference.” Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency (New York, NY, USA), FAccT ’22, June 20, 1375–86. https://doi.org/10.1145/3531146.3533194.

   ↩