How do we build resistance to algorithmic systems of domination when their rewards and punishments are both hidden and immaterial?
While they are composed of humans, mega-machines are fundamentally antihuman, in both their means and ends. “Under the pretext of saving labour,” Mumford writes, “the ultimate end of this technics is to displace life, or rather, to transfer the attributes of life to the machine and the mechanical collective, allowing only so much of the organism to remain as may be controlled and manipulated.” There’s a reason he used the Kremlin, the Pentagon, and General Motors as examples of mega-machines. From their vantage point, the human-cogs have little knowledge of how their actions are incorporated into the functioning of the whole system, let alone choice in the matter. They are treated like automatons or “servomechanisms” who can be kept on a need-to-know basis while work is squeezed out of them.
Now, in the time of networked computing, “smart” technologies, and digital platforms, the image of society as mega-machine is being supplanted by the mega-algorithm—with people acting not merely as cogs but as information nodes, inputs, and outputs incorporated into a calculus of control simply by existing on the network. To accommodate the mega-algorithm, people are atomized by digital technologies and blown apart into streams of data; they are given a unique identity and all their actions—every piece of content generated, every action capable of being tracked—is fed into processors.
Whereas the mega-machine operated by violent means—forcibly divorcing the human-cogs of identity and absorbing their productive and creative energies through wage or slave labor—the mega-algorithm doubles back and promises you reunification with this alienated self through “authentic” (or “creative” or “social”) work completed on your own time. We are sold a desirable narrative about the wealth of networks, decentralized production, cognitive surplus, collaborative consumption, social engagement, and instant convenience. The techno-utopic discourses of emancipation and community that surround the technologies and sociopolitics that make up the mega-algorithm serve as an effective ideological veil, which shrouds the practices of exploitation and control. Don’t think of yourself as an overworked, underpaid laborer trying to hustle for a paycheck. No, you’re actually an entrepreneurial individual, building your personal brand and finding (or making) your niche in the marketplace.
We have escaped the violence of the mega-machine, so the story goes. But the mega-machine and mega-algorithm have more in common than many commentators believe. They are not the same, but nor are they separate; both exist in the same familial lineage, like father and son. As Astra Taylor writes in The People’s Platform, “Many of the problems that plagued our media system before the Internet was widely adopted have carried over into the digital domain—consolidation, centralization, and commercialism—and will continue to shape it.” With both the mega-algorithm and the mega-machine, humans are subsumed into a feedback-driven, information-processing system to which they are at once essential and incidental. The system can’t run without people serving as cogs or nodes, yet in each case, the judgment of those serving the system is rendered superfluous. With the mega-algorithm, functions once assigned to people are assigned to code that becomes one (or many) steps removed from its programmers. The very question of where agency ultimately resides becomes fuzzy, since it is unevenly distributed across different people and times.
The mega-algorithm is powered by a particular idea of how technology fits into political economy: Algorithms serve to hide or obscure the actual human labor involved in business operations. This contributes to a large-scale reshaping of the economics of labor. That is, paying people for the goods and services they produce stops being part of the way business is done and finding jobs that come with security, longevity, and a paycheck becomes more difficult.
Unlike with the mega-machine, it’s not always clear when we’ve been assimilated as a human-node in the mega-algorithm. The systems—like social networks we use to connect with various platforms while hundreds of trackers hoover every piece of (meta)data that can be brokered for the purpose of rating and targeting us—are so interrelated that it appears there’s no escape. It functions rhizomatically: like the roots and shoots of a persistent, massive set of plants, it seems to pop up everywhere. And it is extremely flexible, able to derive value from people in ways that are not obviously exploitative or controlling.
This flexibility, this ability to reconfigure and appropriate additional parts of life extends the mega-algorithm's power further than that of the mega-machine. The technologies of the mega-algorithm disguise human labor by hiding it in plain sight. Every action has the potential to be labor, if it can be processed in the right way. Even the sanctity of our home is not safe from intrusion. As design critic Justin McGuirk argues in an essay on the “smart home,” “the proliferation of smart, connected products will turn the home into a prime data collection node. It is estimated that there will be 50 billion wi-fi-connected devices by 2020, and all of them will collect data that is transmitted to and stored by their manufacturers. In short, the home is becoming a data factory.” This shift makes complete sense in light of the operational logic of the mega-algorithm. Excluding the home from data harvest and monetization would be a foolishly inefficient waste of exploitable labor.
Labor doesn’t become deskilled and cheap by a mega-machine hammering us into cogs—the mega-algorithm prefers stealth assimilation. Consider these examples, two different mega-algorithmic approaches to translation services: Google, for its translation service, uses already available human translations to provide the huge set of linguistic data needed, primarily relying on two giant sources—hundreds of billions of words taken from public-domain United Nations and European Union documents—that they can cheaply use to train their system. If the translators’ copyrights had extended to the types of pattern detection used by statistical machine translation, they could have negotiated for compensation based on the valuable service they inadvertently provided for Google.
Coursera, a for-profit online education startup, also needs translators’ labor, so it can sell its MOOCs (massive open online courses) in other languages. So in April 2014, it created the Global Translator Community. Rather than pay for services, Coursera uses rhetoric of community and solidarity to recruit volunteers to contribute to their “crowd-translating” project. While no money exchanges hands, these “volunteers” must sign a “Translators Agreement” to ensure that all ownership of produced services transfers to Coursera. In a Jacobin article about Coursera’s project, Geoff Shullenberger coined the term “voluntariat” to describe these volunteer laborers. “The voluntariat,” he writes, “performs skilled work that might still command a wage without compensation, allegedly for the sake of the public good, regardless of the fact that it also contributes directly and unambiguously to the profitability of a corporation.”
The voluntariat must in theory derive some kind of fulfillment from this, what Trebor Schulz has called “playbor,” or else they wouldn’t spend free time doing it. But a number of new platforms promise alternatives that allow digital workers to receive compensation, and not just good feelings, for their online free labor. The social-media platform Bubblews explicitly tries to subvert the economics that drive Facebook, allowing users to pay other users directly for content: “they pay a penny per click, per comment, and per like on a user’s posts.” Bubblews pays out in $50 batches, so to earn their first paycheck, users have to rack up 5,000 likes, comments, and/or clicks. One of the company’s founders told Fast Co.Exist, “We started the company because we were reading about how a lot of the social media networks out there were creating the world’s largest unpaid workforces. And we said we could make that different.”
Bubblews has some similarities to Laurel Ptak’s project Wages for Facebook. Though whereas Wages for Facebook is an activist project that serves mainly to politicize the facts of digital labor, Bubblews takes the entrepreneur’s path and promises to cut us in on the profits. As long as we internalize the entrepreneurial ethic, using it to guide every facet of life, we can’t help but be productive human-nodes who keep the data flowing. Your best hope, apparently, is to escape exploitation by becoming the CEO of Your Life, Inc.
The idea that now our social life must constantly be framed in terms of monetization is reductive—serving to further entrench the logics that we should be abandoning and unmaking. It’s telling when the seemingly radical solutions we are presented with—like Bubblews—are still premised on commodification and capital accumulation. How do you fight against a logic or a dispersed and pervasive system that, on a case-by-case basis, doesn’t seem like a big deal and seems to afford a wide range of immediate benefits? What are the rallying points of harm and indignation to motivate action? There’s not even a clear center-mass to aim for. Instead, it can feel like stabbing at a dense fog of abstraction.
To pierce this veil, we must not forget that the mega-machine is still alive and well within the mega-algorithm and is even more insidious. The mega-algorithm's lineage should remind us that its system of control remains a material one, even though it may be implemented with digital tools of capture and through immaterial, intangible rewards and punishments. Overcoming the mega-algorithm's seeming abstract immateriality is the first step toward converting it from an amorphous miasma to an ordinary, mortal monster.