Hannah R. Hopkins

In April 2015, Google released a short video introducing a new platform for cloud computing. Set against soaring instrumental pop music, the advertisement showcases several Google employees — nearly all white men — extolling the technical excellence of Google Actual Cloud, “the world’s first cloud offering running on servers in the troposphere, inside actual clouds.” (00:30 - 00:42) Google Actual Cloud appears as advertised: actual white clouds of data floating above crowds of rapturous onlookers, complete with advanced weather and climate tracking to empower users to “predict storage performance” (01:05) when outdoor conditions get dicey. As a throng of awed Google employees assemble under the shadow of Google Actual Cloud, one of the only two speaking women in the video adds that, “I just love that I can see it.” (01:42) 

Even as the computing industry has proven tremendously profitable for high-wealth communities globally, uneven global ramifications of internet infrastructure still demand more sustained, careful attention. Of course, Google Actual Cloud is an exceptionally well-branded April Fool’s Day joke, but I’m interested to think about why and how it is so effective. Imbricated in the ongoing conversation around the devastating labor and ecological effects of computing infrastructure is the (im)materiality of the cloud, the insistence in popular culture of its weightlessness and disembodiment. As multinational tech corporations engage in concerted public efforts to visibilize the cloud, there remains a need for interventions that address the felt effects of our data engagement. As Safiya Noble argues, “the goal of theorizing a liberatory, intersectional internet is to heighten awareness of how the global communications infrastructure is not just a site of communications affordance, nor is it made equally and equitably available to all people.” (n. p.)

I believe the woman in the video has a point when she says that she loves that she can see “it.” Why don’t we all want to see infrastructure more clearly? What would happen if we understood more closely the contours of data storage and management? In Deep Time of the Media City, Shannon Mattern brings forward a useful and malleable definition of infrastructure, one that takes into account — as historian Paul Edwards argues — that infrastructure “could include hardware, organizations,” and other kinds of latent knowledge or “sociotechnical systems that offer “near-ubiquitous accessibility.”” (Edwards qtd. Mattern, 95). Here, I take up the work of scholars in infrastructure and feminist technology studies to consider: what might change about public understanding of infrastructure if we could “see” what effects our data has on those who are most vulnerable to labor and ecological exploitation?

As Allison Carruth argues, “everyday experiences of the cloud often move online users out of the realm of the sublime and into the realm of the magical; their devices seem to open up conduits into impossible-to-apprehend yet wondrous worlds.” (Carruth 344) While that magic is certainly a benefit for multinational technology firms that rely on “the realm of the magical” to help users forget the materiality of their devices, that realm itself is often devoid of human laborers. In this section, I will focus on Google’s efforts to (in)visibilize the bodies of human workers in one instantiation of the realm of the magical.

In December 2004, Google announced Google Books, a massive digitization project to make millions of books accessible and searchable online. Google Books now hosts more than 20 million volumes online. As Chalmers & Edwards have shown, “Google turned book digitization into an algorithmic process,” (3) a process which attempted to hide the evidence of digitization workers. Regulatory documentation with the University of Michigan, one of the four Google Books library partners, indicates that it was Google’s intention to shield the off-site scanning facilities from outside scrutiny. In this case, “Google prefers that the processes and equipment not be made public.” (University of Michigan Libraries). Chalmers & Edwards describe the contract workers who carried out labor-intensive scanning at Google’s facilities as “light industrial labor, necessary if inefficient elements of an incompletely automated process.” (3) Andrew Norman Wilson, a 2007 Google employee, was shocked when he first realized the cohort of mostly Black and Latinx workers clad in yellow badges he would see each afternoon were actually Google Books digitizers along the fringes of the complex. He writes that his interest began to consume him:

I mined all the information about the yellow badges that I could from Google’s intranet, which led me to the internal name for the team—ScanOps. This class of workers, who left the building much like the industrial proletariat of a bygone era, actually performed the Fordist labor of digitization for Google Books—“scanning” printed matter from the area’s university libraries page by page […] I found some vague meeting notes […] about how they would be excluded from all standard privileges like cafes, bikes, shuttles, and even access to other buildings. […] Why did it seem so secretive?

For Norman Wilson, the distinct yellow ScanOps badge “signified [that the ScanOps employees were"] “not worth the price of integration,”” (Norman Wilson) Yellow-badged contract digitization workers were “directly supervised by the machines they were hired to operate,” (Chalmers & Edwards 3) mostly dispensable but needed to accomodate a wide variation of book types and materials. Norman Wilson’s later project, “Workers Leaving the GooglePlex” attempts to capture the striation and unevenness of Google workers’ status using surveillance camera footage and narration.

Google Books digitization workers’ bodies do appear materially in the content of many Google Books online scans. Website The Art of Google Books collects images of the many Google Books scans that include errors, including scans that feature the hands of Google Books workers. These images are captivating, in large part because they represent an intimacy with the ScanOps workers that by Google’s own design was impossible. Vulnerable to the occupational hazard of repetitive assembly-line style labor, bodies refuse erasure in the Google Books corpus, and the unpredictability of those traces renders its uneven commitment to workers legible. When the librarian in the Google Books 15th anniversary video claims that Google Books “democratized access” (02:56), we must be clear about for whom that democracy is available.

As evidenced by both Norman Wilson’s work and Google’s internal documentation, Google has not long been invested in rendering their workforce visible. Google’s more recent work to demystify data center operations is captured in the 2014 “Inside a Google data center” video (below), wherein several unnamed employees give an overview of their work. At the outset, two White male employees add to a chorus of data center employees: “We keep the lights on. And we enjoy doing it.” (01:09 - 01:11) The only named employee in this advertisement is VP of Data Center Operations Joe Kava, who is similarly insistent that he has a strong affective bond to maintaining Google’s behemoth data center operations.

This representation of data center infrastructure speaks to Google’s interest in populating the representations of networked technologies with enthusiastic, committed employees. Given what we know about labor in this sector, it should come as little surprise that nearly all faces in the video are White.

The advertisement above is listed on the public homepage for Google’s Douglas County, Georgia data center. The link invites visitors to “meet the amazing people behind the technology,” creating the data center rhetorically as a place of life and vibrance.

Navigating further into the page, readers see images and short profiles of data center workers in various positions across Google’s data center sector. Immediately, one is struck by the contrast between the workers in the 2014 video and this image gallery. The imagery is vibrant, almost playful in its depiction of data center infrastructure, again erasing the darker effects of our constant need for data and resultant demand for cheap labor. Rather than engaging the precarious labor of teams like ScanOps or the hidden ramifications of data work visited upon marginalized communities who depend on resources that data centers exploit, Google’s storytelling about itself insists upon a narrative of mutual benefit and even joy in the work of maintaining internet infrastructure. Noble voices a similar call for an intersectional analysis of internet infrastructures and products, “[allowing] for needed linkages between the labor and resources involved in the web and other global communications infrastructure projects that both facilitate, and are a source of, globalized extractive capitalism.” (n. p.) This analysis brings to the fore the tremendous human and ecological weight of Internet infrastructures, and instead centers communities left exploited in the wake of capitalist technological determinism.

A recent effort on Google’s part to make data centers both visible and beautiful has been the 2016 Data Center Mural Project. A promotional video depicts several artists creating large-scale murals on the exterior of data centers, with one artist defining his mission as “[making] something joyful and colorful and try to make people happy.” (01:43 - 01:44) According video’s prefatory material, the murals are meant to make data centers, and in turn internet infrastructures and involved labor legible for a broader public. The narration explains that in Google’s large global data center campuses, “the people and machinery work 24/7 to make things run faster, safer and more efficiently.So much goes in to running the internet inside these buildings, we decided it was time to reflect that on the outside.” (00:24 - 00:40) It is unclear whether such a project’s focus on affect and aesthetics of technological infrastructure does indeed transfer the same positive affect to citizens, observers, or other laborers; but it does call to mind Norman Wilson’s surveillance videos of ScanOps exiting a Google complex. Are they meant to feel joy at the beauty superimposed on infrastructure meant to occlude their very presence?

Set in 2024, Octavia E. Butler’s 1993 ​Parable of the Sower​ follows hyperempathetic Black teen Lauren Oya Olamina as she establishes a new settlement in the wake of climate catastrophe. The novel​ c​ements Butler’s vision of a world ravaged by the white heat of pervasive engagement with personal technologies, and when I visited Butler’s archive at the Huntington Library in January 2019, I was particularly excited to see its first drafts. On one of the earliest drafts, Butler had covered with correction fluid a short line describing how Lauren understands her home’s devastation: of the time just before her ravaged city became uninhabitable, she writes that “it was so easy to do harm.” As we grapple with the capacity and demands of our technological infrastructure, Lauren’s situation is not unfamiliar: our current moment is characterized by precarity in the face of an unstoppable need for faster, bigger data and cheaper, invisible labor.

Allison Carruth argues that our imbrications with data “invite a reflexive ethical stance on how often, how much, and to what ends different individuals connect to and make use of the cloud” (358). Noble’s work — along with the sustained inquiry of Ruha Benjamin, Jessie Daniels, Catherine Knight Steele, Lauren Klein and Catherine D’Ignazio, Donna Haraway, and many other feminist technology scholars — gives us a road map to explore the visible and invisible material dimensions of our data work and data relations. Without such careful thought, it is entirely likely that (in)visibility of the cloud will get in the way of what we can — or even wish to — see of our ever-growing internet infrastructure. A feminist intersectional analysis demands more of our attention than capitalist narratives of unencumbered growth and consumption, and we must take up Carruth’s call for an ethical stance on our own cloud use and policy moving forward.

Carruth, Allison. "The digital cloud and the micropolitics of energy." Public Culture 26.2 (73) (2014): 339-364.

Norman Wilson, Andrew. “The Artist Leaving the Googleplex.” E-Flux. (2016)

Andrew Norman Wilson, Workers Leaving the Googleplex. (2011). http://www.andrewnormanwilson.com/WorkersGoogleplex.html

Mattern, Shannon. "Deep time of media infrastructure." Signal traffic: Critical studies of media infrastructures (2015): 94-112.

Mattern, Shannon. "The City Is Not a Computer." Places Journal (2017).

Noble, Safiya. A Future for Intersectional Black Feminist Technology Studies. Scholar & Feminist Online 13.3-14.1 (2016): 1-8

University of Michigan Library. UM Library/Google digitization partnership FAQ. Available at: http://www. lib.umich.edu/files/services/mdp/faq.pdf (2005)

Here, I want to acknowledge the contours of my own inquiry without jeopardizing my word count for this assignment. I am a white cisgender woman. I live in an affluent nation, where it is possible for me to buy and maintain my own personal technologies with relative ease. Academically and personally, I am actively and purposefully using the platforms which I critique as exploitative. I drafted this paper using Google Drive and my personal Google account. This paper will live on an unlisted page on my personal Squarespace website until after it is graded.

I find this important to mention for a few reasons:

• First, my interest in feminist technology studies requires that I reflect upon my own positionality as a researcher and participant in exploitative systems. 

• Second, my investment in the material ramifications of data culture necessitate that I consider where and by whom the products of my academic life are living.

• Finally, I’m just not satisfied with the ongoing conversations about how data live in the world. Particularly in our capitalist consumption, we’re relentless in our pursuit of tidy supply chain and sustainability narratives that explain where our sweaters were made or how an apple wound up on our kitchen counter. These are material questions, material problems in the world. I think we could expect the same from constant, ambient streams of data that we create and yet do not own.

I hope to convey here that simply knowing where a server is, or “mapping where the data live” as Shannon Mattern argues, doesn’t do the trick of capturing the contours of our data culture. As she writes of urban information landscapes, “It’s also the personnel and paperwork and protocols, the machines and management practices, the conduits and cultural variables that shape terrain with the larger ecology.” (Mattern 2017) In concert with this thinking, I want to insist fervently that intersectional feminist technology studies teaches us that we must understand how our data comes to be in the world, including and especially the labor practices that make it possible for me to send you this very writing assignment. The exploitation that underpins our technoculture is always already present in our digital marks. We are all implicated in the phenomena that comprise digital production, and our responsibility to address exploitation of marginalized workers does not decrease at scale.