The Social Industry

Abstract

Historical mass media and contemporary social media are typically seen as opposites. “The culture industry” was the term used by the Frankfurt School in the 1940s to explain the emerging commercial mass media. The culture industry was portrayed as a semi-fascist apparatus of indoctrination. It selected cultural products and made them popular based on obscure determinations of economic value. In contrast, the common view of contemporary social media is that it is more democratic. Using voting algorithms and human voting, social media can finally realize widespread participation that was denied to the passive audiences of the mass media system. Social media appear to remove the bottleneck of the mass media system, allowing everyone to aspire to celebrity, or at least popularity. However, despite these appearances, social media have also now evolved into an elaborate system that selects social products and makes them popular based on obscure determinations of economic value. Social media platforms filter, censor, control, and train—and they may do so without the user’s awareness. Advances in computation now make a social media industry possible that is based on individual difference and action rather than sameness and passivity. But in other respects, the social industry resembles the culture industry: the co-option of culture has been superseded by the co-option of sociality. The word “social” may then be the biggest challenge facing those who study social media. Our task is to rescue genuine sociality from the emerging social industry.

The culture industry was so named by the refugees of the Frankfurt School who explained it with references to Joseph Goebbels and propaganda. ¹ In these circumstances, it was natural that the historical media industries felt authoritarian. When Google founders Larry Page and Sergei Brin first described the PageRank technology behind the Google search engine, they wrote that it was “democratic” (Page et al., 1999, p. 11), and the metaphor of democracy made sense as a contrast. Just as it is sometimes said that anyone can grow up to be president, they noted that any Web page on the Internet could in theory be returned as a result to a Google search. Under Google’s PageRank algorithm, the “winning” Web page that was offered to the user first rose to the top because other Web pages linked to it, a decentralized decision process sometimes called “voting” in computer science (e.g., Lifantsev, 2000). In the old media systems of movies, news, music, and television, despotic media executives used mysterious processes to make decisions on your behalf. New media were the Allied tanks rolling through Paris; they were the fall of fascism.

“Social” media presented itself as taking the democratic liberation further: a dramatic expansion of the franchise. In the earliest days of social media, it was thrilling that the Internet was “enabling conversations among human beings that were simply not possible in the era of mass media” (Levine et al., 1999, p. xxii). Platforms took the affective gemeinschaft of personal conversation and intermingled it with the gesellschaft—the transactional, rational production of the culture industry—in new ways (Tönnies, 1887/2001). Web links were once analogized by Google as “votes,” but with the advent of explicit rating systems, now actual votes were votes (upvotes, likes, favorites). As anyone could contribute to the hopper of social media content, it seemed like anyone could be a celebrity, or at least get a lot more attention than those in the dark days of the mass media.

There have been other effects of the “social” reshaping of the media system. Our lives are less private, we are encouraged to develop a personal brand, and there is a tendency to “present . . . oneself as a celebrity regardless of who is paying attention” (Marwick, 2013, p. 114). But these consequences flow from a new meritocratic assumption that we are all participating in a meaningful way. This status update didn’t get any likes, but maybe the next one will. The next popular meme could be me. When I post and upload to social media, I’m talking to people. We act as though the culture industry bottleneck has disappeared. The dead hand of Hollywood studio executives has nothing to do with us.

The Culture Industry Has Been Reinvented

Instead of overturning the culture industry, we have reinvented it. Instead of social media being the opposite of the mass media, it is its extension (McChesney, 2013). We ran from the industrialization of cultural products only to industrialize our personal conversations and intimate moments. On contemporary social media platforms, despotic algorithms use mysterious processes to make decisions on your behalf. Social media popularity does not guarantee that the popular social media object has any intrinsic value. And once something becomes popular on social media, it can sometimes remain popular on the basis of its familiarity alone.

Savvy audiences always knew that “economic selection” by the culture industry never guaranteed merit (Horkheimer & Adorno, 1944/2002, p. 98). If a motion picture actor was famous and seen everywhere one decade, disappeared the next, then returned for a comeback in the third, no one would seriously claim that the actor’s decline was due to quality. Had his acting skills decayed for 10 years, then strangely returned? Certainly not. Any variations were due to ineffable processes of selection and taste. Popularity did not guarantee that the popular media object had any intrinsic value. And once something became popular, it could sometimes remain popular on the basis of its familiarity alone (Boorstin, 1961). Still, we persist in thinking that popularity on social media involves merit, despite the mounting evidence to the contrary.

Consider social media sites like Reddit and Imgur. These have millions of unique users who upload billions of items of content. Personal stories, memes, and pictures of cats are filtered through a voting process that is intended to direct other users toward the best content (Gilbert, 2013). Most visitors see only the content that has already been heavily upvoted—among experts this is termed “social navigation.” Yet, veterans of the site discovered via trial-and-error experiments that their posts fail and succeed almost randomly. Exactly the same content submitted at two different times might “lose” with zero votes, then “win” with thousands of votes. Although the platforms intend to reward interesting original material, identical uploads recirculate through the system again and again—popular “original” posts often turn out to be taken directly from other users. Voting brings with it both the promise and the horror of democracy. Upvotes privilege inoffensive, bland content. Users copy the votes that they see other people making. Active participation is rare, leading an engaged few to decide for the apathetic many. And “original” social media content is quite often an excerpt of a media product produced by the culture industry. The dominant narrative of technological progress has some difficulty dealing with reversals (Streeter, 2010), so we choose not to see any steps backwards. Have the fascists returned?

Democracy Is the Wrong Metaphor

The more we investigate “social” media, the more it is clear that democracy has been the wrong metaphor all along. When the word “voting” is used as a metaphor in the descriptions of algorithms written by computer scientists, this is often nothing like voting. The common sense formulation of voting when applied to social media might be that anyone can contribute content (active suffrage) and anyone may vote (universal suffrage), but all votes must be weighted equally (equal suffrage). This does not begin to capture the process of algorithmic content selection. Even in Google’s earliest version of PageRank, “votes” were made by Web pages, not people, and many disenfranchised pages were considered not worth counting (Introna & Nissenbaum, 2000). In contemporary social media, it appears that the cat picture with the most upvotes will prevail, but what we see is actually subjected to a secret process that determines its relevance—ultimately judging whether something will be shown at all (Gillespie, 2012). In a recent study, the majority of social media users were not even aware that any filtering was occurring (Eslami et al., in press).

When I type a status update on my Facebook account, I may think that I am writing to my friends in public, but in fact Facebook evaluates what I write and may decide not to show my post to anyone. A bottleneck still exists. In a recent Washington Post experiment, about 60% of all posts from friends were not shown on a reporter’s news feed (Dewey, 2015). Social media content may appear or disappear according to reasoning that is unrecoverable. Facebook revealed that it experimented with hiding status updates on the basis of how many happy or sad words they contained (Kramer et al., 2014). Social media platforms like Facebook and Pinterest also maintain an army of content censors who have banned images of breastfeeding, scientific diagrams, famous artworks, and political speech, sometimes acting at the request of dictatorships (The Economist, 2014).

Rescue Sociality From the Social Industry

Years after coining “the culture industry,” Adorno (1967/1975) reflected that the word “mass” was excluded from the phrase because it was a dangerous word. The word “mass” might give the reader the wrong impression: that the culture industry contained “something like a culture that arises spontaneously from the masses themselves” (p. 12). In the culture industry, the audience may select programs, buy tickets, or change channels, but in doing so, each person acts as “an object of calculation; an appendage of the machinery” (p. 12). While some misguided analysts might see this as “even democratic, since it responds to a demand” (p. 16), in fact the demands are produced and ultimately shaped by the industrial apparatus itself. The culture industry is a sham democracy that solicits input only for its own ends. The culture industry produces the audience; it is not governed by them.

In a similar manner to the word “mass,” the word “social” may be the biggest challenge facing those who study social media. Horkheimer and Adorno feared the culture industry for many reasons, but one was its production of sameness and passivity. Human ingenuity has addressed this flaw, and the new social industry is outwardly organized around difference and action. The word “social” was intended to signify subjectivity, interactivity, and participation. Surely within these behaviors, there is a genuine human sociality, but what has been built around it is a calculation machine of dizzying complexity. The machine implements obscure determinations of value (Pasquale, 2014), with users acting as its input, output, or gears.

Andy Warhol once produced an uncanny premonition of social media. He explained Pop Art in a 1963 interview by saying to the interviewer that “I think everybody should be a machine. I think everybody should like everybody.”

The interviewer asked, “And liking things is like being a machine?”

To this Warhol replied, “Yes, because you do the same thing every time. You do it over and over again” (Honan, 2014). For the users of social media, the most pressing job is not to pretend to “vote.” It is not to like things over and over again. Instead, it is to work together to rescue genuine sociality, community, and democracy from the proliferating industrial machinery that has devoured it.

For the Frankfurt School, the culture industry was dangerous because it was soporific. It conveyed ideology without appearing to do so, and thus it replaced a spontaneous culture with something manufactured. Social media are so far tracking a similar trajectory. The danger of handing our sociality to the social industry is that we may find it impossible to see what we have lost by doing so. There may be truly emancipatory and positive relationships that computation can make possible. In this hypothetical future, we would produce algorithms that are actually social and are actually democratic. And this would not be a metaphor.