Electronics manufacturing is toxic and disabling. Some of its most crucial tech manufacturing tasks are still done by hand. Undergig labor in electronics production is intimate, often metal to skin. (And yet, paradoxically, it is the human that is rendered toxic to the tech: key manufacturing steps for electronic devices have to be executed in clean rooms, with workers shrouded in protective garb.) Though automation has taken over some of this work, it has not taken over all of it; the scale of demand for circuits and the devices that use them keeps pace with the need for human bodies and human capital to fuel the economy of anticipation, growth, and expansion.
Even the “good” digital jobs (those in the tech industry that are non-toxic, well-paid, with the possibility of advancement) are still precarious; like the good life, good jobs now have a phantasmatic quality defined by frequent shuffling and layoffs. And these are the jobs that people struggle to get and keep. The worker’s body pays the price; stress, disease, mental health all take their toll. The depletion economy produces massive amounts of disability all along its circuits. It also produces vast amounts of toxic detritus.
More than the electronics themselves, toxicity is the tech economy’s biggest export. This is why we turn to how toxicity – the spread of environmental harm and vulnerability in the depletion economy – is the condition for digital production today.
Toxicity operates as a metaphor. Dominant video game cultures propagate “toxic masculinity.” Abusive CEOs at start-ups perpetuate tech’s “toxic” workplace culture. Properties held by struggling financial institutions might be “toxic assets.”1 These rhetorical uses signal how the idea of bodily invasion and infection weaves in and out of our contemporary technological discourse, even as digital technology purports to be disembodied, smooth, clean.
Yet it is more than just a metaphor. Toxicity burrows deep in flesh as people breathe in its evanescent particles. Brett Walker’s deep ethnographic work on heavy-metal poisoning in Japan teaches us that industrial toxins have no boundaries – their traces can be found in deaths from insecticide contamination, poisonings from copper, zinc, and lead mining, or in the congenital deformities that result from methylmercury factory effluents.2 They all precede and underwrite the age of digital dominance.
What could it mean to think about toxicity as both the production of disabled bodies and as a potent figure for understanding the subject? The purging of toxicity is imagined as a way to re-establish the purity of the subject. Just as toxicity is everywhere, so too are attempts to purge the body of it: juice cleanses, water cleanses, herbal cleanses, colon cleanses, digital cleanses. But as Alexis Shotwell reminds us in Against Purity, there is no easy way of immunizing ourselves against our impure pasts and complicit presents.
Many of us are settlers living on unceded native land, stolen through genocidal colonial practices. We feed domestic animals more food than starving people lack, and spend money on the medical needs of pets while eating factory farmed meat and spraying our lawn with pesticides that produce cancer in domestic animals […] We cannot look directly at the past because we cannot imagine what it would mean to live responsibly toward it. We yearn for different futures, but we can’t imagine how to get there from here. We’re hypocrites maybe, but that derogation doesn’t encompass the nature of the problem that complexity poses for us. The “we” in each of these cases shifts, and complicity carries differential weight with our social position—people benefiting from globalized inequality are for the most part the “we” in this paragraph. People are not equally responsible or capable, and are not equally called to respond.3
There is no cleanse for precarity or for our particular roles in sustaining it.
It’s no secret that the purchasing of digital devices funds worker and human rights atrocities. Our devices rely on labor and materials that support structures of exploitation and violence. Each step in the production process exposes workers to a different form of toxicity.
It begins with instrumentalizing the periodic table; rare metals are essential materials for ever smaller and more powerful digital devices. The map of rare metals changes the geopolitics of where mining happens as nations scramble to control the extraction of crucial materials. New metal mining industries graft onto formerly colonized landscapes (Latin America, Africa, Australia); they engage the bodies of miners exposed to these metals as countries race each other to control these growing markets. The production of electronics, moreover, requires the mining of high-value raw minerals – gold and the “3Ts”: tungsten, tin, and tantalum. Digital devices on the shelves of your closest Amazon warehouse or sitting comfortably in your pocket are possible thanks to the entanglements between metal mining industries and the enduring detritus of imperial refuse.
Next, we might think about how the assembly of devices also endangers workers by exposing them to toxins. Supply chains and the global assembly line converge on Asian (often women’s) bodies as they assemble toxic components of high-demand devices. Their reproductive, life-giving capacities pay the price for ubiquitous electronics. One of the richest companies in the world, Apple, has the highest industry mark-ups, made possible by the labor of workers who get sick, lose the capacity to bear children or have borne children with serious disabilities after working in their factories. The depletion economy exports toxicity to import cheaper devices to parts of the world privileged enough to purchase them.
Recent projects of upgrading the manufacturing industries in the coastal regions of China are aimed at taking human inefficiency out of the loop. While China is busy upgrading its factories into high-tech plants, the US strives to “bring back” the slogan “Made in America.” US industrial towns are importing toxicity. Under pressure for reelection, politicians are calling for a return to a manufacturing economy. In 2017, the city of Janesville, Wisconsin campaigned hard to attract a Foxconn factory to its small town to replace a shuttered Ford plant that had employed unionized workers. This plant was never built, but had it succeeded, the kinds of jobs that these workers would have had would have exposed them to toxicity. This promise of a return to “Made in America” sits side-by-side with the desolated integrated circuit factories of California, left behind as vast Superfund sites.4
After consumers purchase and use electronics, they interact with platforms such as Facebook, which in turn rely on low-cost, vulnerable labor to perform the chronic and grisly task of content moderation. The labor of content moderators involves having to watch and remove toxic content – sexually graphic and violent material (images or references to pedophilia, necrophilia, animal abuse, beheadings, suicides, murders, etc.). Facebook’s basic moderation is typically outsourced to countries like the Philippines and India for their familiarity with Anglo-cultural norms as a result of a history of colonization and their willingness to accept extremely low wages (on the order of $1 to $2.50/hr).
Machine learning algorithms designed to detect violent, illegal, inappropriate, and disturbing content online do not simply and automatically remove media containing child pornography from the internet and keep moving. Instead, such media are sent to content moderators who determine whether or not flagged content did in fact contain instances of child pornography. Social networking services make poor and poorly paid workers in the Philippines and in India moderate violent or graphic content online.5 They often do so to the detriment of their own mental and emotional health, so our timelines and feeds can remain relatively innocuous. This kind of unregulated and non-unionized work is not foreign to workers in the US. More than 5 percent of workers there rely on this kind of crowd-work from tech companies.6
As Mitali Thakor has argued, the incorporation of child-abuse detection algorithms by law enforcement agencies has created a new hybrid machine-labor ecosystem that is not limited to the work of traditional law enforcement officers, but more comprehensively includes algorithms and the computers that run them, computer scientists and programmers, and content moderators in addition to law enforcement.7 Policing institutions justify marshaling resources for these hybrid machine-labor ecosystems of “algorithmic detectives” through a racially exclusionary appeal to child innocence in which innocence is always conferred to potential victims in photographs through a subtle, almost unconscious evaluation of their proximity to the figure of an ideal white child.8
It is tempting to think of all this in terms of an unequal relationship between the global north, as exporter of toxicity, and the global south, to which it is exported. This would be an over-simplification, but it cannot be denied that surplused populations across the globe are often the most precarious test subjects that serve the depletion economy. And the cardinal split of the globe doesn’t quite cut, as the proliferation of toxicity in depletion zones attests.
To examine precarity in our contemporary moment is to be attuned to the damages, the stubborn remainders (or reminders) of modes of power invested in the differential management of human and nonhuman lifeworlds.
Let us not overlook how toxicity is also dumped on people of color and other precarious populations internally, within the territories of the global north as well. Such territories and their populations can become laboratories for testing the results of toxic procedures. They are close enough to study, precarious enough to lack the power to object, and yet held at arm’s length, internally, in the belly of the supposedly protected global north.
Consider, for instance, that a 1970 study of the correlation between birth defects and radiation, specifically from uranium mining among Shiprock Navajo workers, found that the “association between adverse pregnancy outcome and exposure to radiation were weak,” but that “birth defects increased significantly when either parent worked in the Shiprock electronics assembly plant.”9 Similar correlations were found at other assembly plants in California and elsewhere. Epidemiologists knew what the industry didn’t want to know: the suffering of indigenous women and babies was part and parcel of this industry.
The laboratory is where people of color, indigenous people, and poor people are. The reservation in the United States is a space where these three identities live together. It is where experiments have been conducted for more than two centuries now. Shiprock’s Fairchild plant and others like it were a space for multiple kinds of experimentation on women of color; there was also a uranium mine nearby, and a power plant, operated by Kerr McGee (Karen Silkwood, a white woman who blew the whistle on this toxic industry, suffered from serious organ contamination after working at Kerr McGee. She died in a mysterious car crash after suing the company).
Thus, a Navajo woman who worked at the Fairchild Semiconductor Corporation’s electronics plant had a higher chance of bearing a child with birth defect(s) than if she was exposed to radiation. But her chances of this occurring if she worked manufacturing semiconductors were even higher than that.
It was legal to export toxicity to Navajo women because the plant was built on Native land – because as sovereign nations, the Navajo were not considered part of the US, not subject to the same laws and protections. They were forced to receive this import, in exchange for the export of the tiny components that would power rockets, satellites, calculators, and eventually computers. The State of New Mexico as a whole has high levels of both toxicity and poverty. Some of these poisons were, like uranium, “native,” and drove the siting of national labs like Sandia and Los Alamos. New Mexico was a space that technologists, experts, entrepreneurs, the military, and politicians imagined as empty,10 a place where weapons could be tested and new kinds of labor could be prototyped.
The semiconductor industry in the US first knew about the effects of toxic manufacturing practices on female workers in 1984, when a graduate student moonlighting as a health and safety officer at Digital Equipment Corp told a young new assistant professor, Harris Pastides, that women who worked at these plants experienced extremely high rates of miscarriage. Digital Equipment Corp “agreed to pay for a study” that proved that this was true; three subsequent studies confirmed it, and the results were reported to the Semiconductor Industry Association, which ignored it.11
A photo essay commissioned by Bloomberg titled “These Women Are Paying the Price for Our Digital World” shows Korean women who suffer from brain tumors, cancer, and other disabilities as a result of their work at a massive Samsung plant.12 Their work is the foundation for South Korea’s identity as a high-tech nation. The precarity experienced by these Korean women had already happened on Native land in the US, almost ten years earlier.
Reservations have always been economic laboratories, of cigarette consumption and gambling, things we call vices, simply because these products were not regulated there. These areas were made into experimental sites for the digital. From 1965 to 1975, 20 years earlier than the Pastides and other studies, almost 1,000 women worked at the state-of-the-art Fairchild Semiconductor plant in New Mexico, on Native land. These women, along with the thousands of others working in the Fairchild plants in Asia and the US, built the digital industries.
It was the precarity of indigenous women (who all lost their jobs when the plant was taken over by American Indian Movement activists) that created the conditions of precarious workers in the Bay Area. Like their indigenous sisters, women in the Bay Area suffer far more breast cancer than the norm; no one seems to have nailed down the cause.
Silicon Valley exports precarity to places such as Shiprock, New Mexico, the Philippines, Korea, and Malaysia and because the industry is built upon precarity; it shifts locale to where labor is the cheapest and least accountable to regulation. Indigenous women’s precarity produces every other kind of precarity in the digital industries. The number of people sleeping in cars, in tents, in RVs – more likely people of color but also the poor – on sites such as the Stanford Campus’s El Camino Real, are an eloquent testament to the impossibility of living with dignity in the Bay Area, where real estate is unattainable except by the wealthy, but where jobs as contractors and freelancers are to be had.
The body is a lab for precarious living; precarious bodies are the crash test subjects for the juggernaut of extraction, leaving behind on its trail sites, land, bodies marked by toxicity. Toxicity is never separable from the question of who or what is credited as human.13