Fixing to Die

What would an “abolish not reform” approach to tech look like?

Full-text audio version of this essay.

In a recent piece for the New Yorker, Keeanga-Yamahtta Taylor described the police violence that came in the waning days of Derek Chauvin’s trial for the murder of George Floyd. As some celebrated the apparent resolution of one atrocity, other atrocities were just beginning the all-too-familiar cycle: Documented violence, an attempted cover-up, public reckoning, legal proceedings, a verdict of some kind, and then another new wave of “incidents” by law enforcement.

Tech companies have also grown into a certain kind of cycle, with their products regularly generating widespread social harm. Facebook, for instance, has provoked such atrocities as genocide, fascist insurrection, and mass shootings. Just as it was publishing transparency reports and apologies about its role supporting interference in the 2016 election, it helped enable the Myanmar military’s genocidal violence that killed 25,000 people and displaced 700,000 more. Last summer, a Facebook post played a key role in the Kenosha, Wisconsin, shooting, and Mark Zuckerberg acknowledged the mistake and apologized. Then, during the 2020 election season, Facebook helped facilitate the “stop the steal” conspiracy theory and allowed Trump supporters to organize the violent attack on the Capitol, all while also allowing disinformation denying the ongoing Uyghur genocide in China to spread.

Whatever the next egregious tech abuse will be, the response will likely follow the established pattern: After public criticism, companies will apologize, perhaps testify before Congress, share their plans to improve, and then move on to provoking a new wave of atrocities. Without serious intervention, the cycle will likely continue.

Faced with cycles of police violence, a growing movement for police abolition seeks to reimagine public safety from the ground up. For abolitionists, police violence is best understood as an everyday symptom of policing rather than a set of anomalous incidents, requiring not mere reform but an entirely new paradigm for justice and community safety. The emerging abolitionist movement offers important lessons for breaking cycles of harm, imagining new futures, and delivering meaningful change.

Whatever the next egregious tech abuse will be, the response will likely follow an established pattern. Without serious intervention, the cycle will continue

What if the critique of tech companies followed similar principles? In some ways, the limits of reformist efforts to tech are approaching fast. Labor organizing can prevent or delay government contracts, regulation like GDPR and CCPA can help protect user privacy, and five ambitious antitrust bills in U.S. Congress may even break up some tech companies in the near future. These bills would also address monopolistic practices, forcing companies to pay higher merger fees and offer portability options to simplify migration to other products — all welcome changes.

But just as police reform cannot change the fundamental goals of policing, tech reform cannot change the fundamental goals of large technology companies. What would it look like to apply an abolitionist approach to big tech?

One of the key rhetorical shifts in the police-abolition movement involves focusing on deeply rooted everyday harm rather than the “itemized atrocities” that can dominate public discourse. Lists of atrocities can sometimes be helpful for hooking attention, substantiating historical analysis, and working toward public accountability, but abolitionists argue that real change requires close attention to what Saidiya Hartman, in Scenes of Subjection, calls “the terror of the mundane and quotidian.” Discussing the concept, Tamara Nopper and Mariame Kaba note that long lists of shocking incidents can lead to desensitization, as “the atrocities itemized need to happen more often or get worse, to become more atrocious each round in hopes of being registered.”

The “incident” framing also suggests that shocking events are anomalous problems of excess, as if a “bad apple” police officer merely knelt for too long or shot too often. As Nopper and Kaba argue, “we must accept that the ordinary is fair, for an extreme to be the problem.” But an abolitionist politics suggests that the ordinary is anything but fair.

After all, many police forces were not originally created to serve and protect the public interest. In a piece for the New Yorker, historian Jill Lepore traces “the invention of policing” back to urban slave patrols. In a similar vein, historian Simon Balto describes how the early Chicago police department was “developed primarily by elite business owners in the city with the primary purpose of controlling immigrant behavior.” The early Chicago police department supported Al Capone’s organized crime group, and squelched groups working for labor rights and tenant rights.

If the problem is policing itself, then reforms, however useful they might be in reducing immediate harms, will result in only reformed cycles of harm

Within a historical framework, “incidents” of police violence are not really incidents at all, but inevitable events that serve the original purpose of the larger criminal punishment system in the U.S. As legal scholar Dorothy E. Roberts puts it, that purpose is “reinstating the subjugated status of Black people and preserving a racial capitalist power structure.” Historical accounts show that from the beginning, the system of policing guaranteed everyday violence of surveillance, harassment, and abuse targeted toward Black people.

If the problem were one of excessive violence, then it could be solved with banned chokeholds, more body cameras, and tweaked police-training programs. But if the problem is — and always has been — policing itself, then reforms, however useful they might be in reducing immediate harms, will result in only reformed cycles of harm. In the words of Naomi Murakawa, author of The First Civil Right: How Liberals Built Prison America, reforms merely “tinker with the techniques of police violence.” Any real solution, then, would need to imagine alternative institutions and alternative paradigms for public safety.

In seeking to borrow from police and prison abolitionist theory to address the harms caused by technology companies, a helpful starting point is to distinguish between listing big tech’s “itemized atrocities” and articulating its equivalent of the “terror of the mundane.” Many academic articles, essays, and documentaries include different versions of the “itemized atrocities” list, often including “incidents” such as 2016 election interference or Facebook’s role in the January 6 attack on the Capitol. Other example “incidents” include Google’s role in radicalizing the white supremacist responsible for the Charleston church massacre, Amazon selling fraudulent “coronavirus supplements” during the pandemic, and Uber’s “mistakes” calculating low wages for workers.

If these events are anomalous, then incremental reforms such as expanded privacy features, more content moderators, improved AI systems, and new antitrust bills might prevent future incidents with technology companies.

But just as a growing chorus of abolitionists recognize police violence as “the system working as designed,” a growing chorus of tech critics view the atrocious “incidents” committed by tech companies as inevitable consequences of the products and services being offered. The incidents come to appear as familiar, rather than exceptional.

As with police violence, the itemized atrocities can be helpful insofar as they illustrate clear cycles of harm. That is, the growing list of incidents suggests that they may not be “incidents” at all. Chris Gilliard’s metaphor of technology companies as polluting factories further illustrates the point: Harm flows continuously and inevitably from the design of certain tech products. Uber was premised on low-wage gig labor, Twitter was premised on reactive dialogue, and Amazon Ring was premised on invasive surveillance. The associated consequences are simply unavoidable, and in many ways actively stimulated.

To use Gilliard’s words, Facebook’s business model necessitates a continual flow of hateful and otherwise problematic content. As articulated by the “Stop Hate for Profit” campaign, such content drives engagement and can help Facebook make advertising money. But even if Facebook removed advertising and became a nonprofit organization (as in a recent thought experiment from Casey Newton), the flow would not stop, because there is no way of effectively moderating content at Facebook’s scale. With tens of thousands of workers reviewing hundreds of posts per day, Facebook can still only review a negligible fraction of content across its products. To make matters worse, hate speech and conspiracy theories proliferate further in the private groups that Facebook has been promoting for several years.

Low-wage labor, reactive public dialogue, invasive surveillance, and problematic content are some of the everyday harms perpetrated by large tech companies. These harms sometimes escalate to high-profile atrocities, but every day they are part and parcel of products offered by the likes of Uber, Twitter, Amazon, and Facebook.

A growing chorus of tech critics view the atrocious “incidents” committed by tech companies as inevitable consequences of the products and services being offered

Google also perpetrates everyday harms in its course of doing normal business. Safiya Noble showed in 2013 that Google search results for “Black girls” reflected discriminatory stereotypes on the basis of race and sex. While the company has since plucked out some of the most egregious examples of algorithmic racism and misogyny on its products, The Markup found last year that Google’s advertising systems still associated “Black girls” with pornographic content. This year, an extensive analysis by Rodrigo Menegat found similar problematic stereotypes in Google’s image search results. Even Google’s response admitted that stereotypes embedded in their algorithms present an endless game of whack-a-mole: “Our improvements will not solve every possible query in every country or language.”

Viewed through an “incident” framing, election interference, hate speech, conspiracy theories, and algorithmic racism were all just anomalous, maybe even resulting from “bad actors” abusing tech products. But an abolitionist lens recognizes the cyclic nature of these atrocities and the fact that many tech products support this “misuse” by design. Mar Hicks sums it up with a technical metaphor: Racism and misogyny are features, not bugs, of how large technology companies profit.

A number of helpful frameworks help formalize the everyday harms associated with large technology companies. For Shoshana Zuboff, the key problem is that these companies apply capitalist market incentives to the practice of surveillance. For Ethan Zuckerman, the “original sin” of the internet was its reliance on advertising, which made constant manipulation the “default model” for funding online infrastructure. For Nick Couldry and Ulises A. Mejias, the root problem is dispossession and extraction, the fact that large technology companies extend settler colonialism into the digital age through
“data colonialism.” These and other frameworks have helped drive the “techlash” of recent years, as each provides unique insights about the data capitalist complex — the system that seeks to monetize any human behavior involving technology. The data capitalist complex also represents an underlying ideology, one that paints humanity as Homo economicus, existing primarily for the marketplace.

Like the prison industrial complex, the data capitalist complex has its alibis. Facebook’s executives and employees regularly tout a “net positive” impact on the world, and Google often showcases technology to address climate and health care challenges. Besides, shouldn’t we just be grateful to have internet search and social networking “at no cost?” For all the harm they do in the world, at least Google and Facebook never charged you $1 a month (more than enough to cover either company’s revenue, based on a 2013 analysis). And if people actually prefer monthly fees over advertising, surely “the market will do its thing” and eventually offer a monthly fee option. The best options will prevail, we just need “fair competition.” Then, instead of griping, anyone who wants Amazon to treat their workers better should simply choose another way to shop online, and anyone who wants the same from Uber should simply find another ride-hail app. Once there are more options, consumer choice will help the “invisible hand” to sort things out.

The antitrust bills in Congress promise to prime the marketplace for such options, to “level the playing field” so that consumers can more easily use different online shops, ride-hail apps, social networks, and search engines. But leveling the playing field and changing the rules will not uproot the underlying game, and thus will not address the ultimate culprit. If the game is still capitalism, the data capitalist complex will only manifest in new varieties. Companies are already offering samples of this reform: Ring surveillance, now with encryption! Reactionary dialogue on Twitter, now for crowdsourced content moderation! Google advertising, soon targeting to small groups instead of individuals!

In a world where Congress passes the currently proposed antitrust bills, these reforms will likely continue, unleashing a new wave of privacy-washed, green-washed, pink-washed, and ethics-washed tech products. This kind of reform does not change the underlying assumption that human behavior — especially human behavior involving technology — aligns with the end goals of capitalism. So in addition to breaking up big companies and creating new ones, an abolitionist movement for big tech must also offer alternative ideologies.

To counter the entrenched ideology of punitive justice, police abolitionists offer transformative justice. This alternative paradigm can be loosely defined as a collective approach to “making things right” which seeks to avoid violent state systems (like police and prisons) and take active measures to prevent violence (like building community and resolving conflicts). In parallel, a helpful paradigm shift for tech abolitionists might be from capitalist competition to mutual aid. Mutual aid can offer both practical and ideological alternatives to big tech, instilling different kinds of everyday interactions with each other and with technology, while also establishing different goals for the future.

Rather than framing life in terms of hierarchy and competition, the theory of mutual aid (in anarchist Peter Kropotkin’s work) frames life in terms of interaction and cooperation. Discussing his book on the topic, Dean Spade contrasts mutual aid with capitalist charity: Whereas charity often serves “to quell uprisings that people would engage in against systems that are so extractive,” mutual aid “focuses on helping people get what they need right now, as we work to get to the root causes of these problems.” Through the lens of mutual aid, the world is fundamentally abundant. Rather than competing to accumulate the most resources, one of the main tasks in life is collaborating to make sure others have the resources they need.

Whether working toward an abolitionist future based on transformative justice or mutualistic technology, some of the most important work is imagining what it might actually look like. In widely circulated comments from last summer, Congresswoman Alexandria Ocasio-Cortez pointed out that a world without policing may not require much imagination at all: Wealthy suburban communities tend to prioritize funding for education, youth, health, and housing projects over funding for police. And when someone does something harmful in a suburb, the community will “bend over backward to find alternatives to incarceration for their loved ones to ‘protect their future.’” As Kaba puts it in this interview, “there are groups of people who are living a type of abolition now … think of [affluent, white] neighborhoods in the Chicago area like Naperville where there are no cops to be found.”

Whereas capitalism presumes individuals can choose the best tools from the marketplace, mutualism presumes communities can create the best tools for themselves

Similarly, there may already be communities living in a world without big tech, at least to some extent. Many wealthy executives in the technology industry abstain from using big tech products, and force their children to do the same. But neither families in the suburbs nor tech executives from The Social Dilemma offer truly viable alternatives, because they both fail to grapple with the roots of today’s issues. When it comes to children, for example, suggesting abstention as a solution can “gloss over how young people can make positive social connections online,” and also neglects larger systemic factors that impact childhood wellness, such as economic anxiety and the climate crisis.

For meaningful change to take place, a world without big tech products must also be a world without the data capitalist complex and its supporting paradigms. In part, this means there can be no wholesale replacements that work on the same massive scale as big tech — communities must imagine and develop their own local alternatives for online information portals, social networks, and other tools that operate as monopolies in today’s ecosystem. Whereas capitalism presumes individuals can choose the best tools from the marketplace, mutualism presumes communities can create the best tools for themselves.

Mutualistic information portals are closer than one might think, at least in physical form. Since the 1700s, local public libraries have been storing, curating, and distributing information in the public interest. They also provide a host of other public services to their respective communities, including physical meeting space, educational workshops, public internet access, and support for new immigrants. A future without Google might be a future where search engines are modeled after libraries, as suggested in Safiya Noble’s work. Sure, some people might still need massive commercial search engines for some use cases, but the ecosystem could operate locally and mutualistically, without presumptions of cut-throat competition, unpredictable changes, and centralized control.

In terms of social networks, a few projects provide helpful glimpses of a mutualistic future. The app Herd, for example, has been designed to “cultivate a kinder, friendlier, calmer environment. Herd’s founders said they would “rather make a platform that means a lot to a smaller group than nothing to millions.” Within this paradigm, Herd is simply designed to support the people who use it, not to grow and profit by any means necessary. This is a promising shift for a future with localized, mutualistic technology, and these ideals could also help guide similar projects.

With 20 employees and 200,000 users, Front Porch Forum — a once-a-day newsletter that connects neighbors in Vermont — is another platform “that means a lot to a smaller group.” Co-founder Michael Wood-Lewis has said there are no real plans to “scale up” in terms of geography, features, or otherwise. Because of its small scale, Front Porch Forum is totally useless to big tech companies, but all the more meaningful to people who actually use the app every day.

If and when Congress passes antitrust bills to reform large technology companies, that must be seen as a beginning. The bills do not guarantee transformation. Just as it is impossible to separate policing from the punitive logics of the prison industrial complex, it may prove impossible to separate Google, Facebook, Amazon, and Uber from the extractive logics of the data capitalist complex. To abolish big tech will require abolishing its core capitalist ideologies, allowing communities to build and maintain alternative technologies premised on care, cooperation, and mutual aid.

Jack Bandy is a PhD candidate at Northwestern University. His research and writing focuses on public interest computing, and explores the social impact of algorithmic platforms like Twitter, Apple News, and TikTok.