The bias of code: Why Facebook’s trending topics scandal is beside the point

It’s fascinating that just a couple of months since posting about social media’s emerging bias, Facebook becomes embroiled in scandal related to political bias in it’s trending news curation process.

Sen. John Thune’s (R-S.D.) letter to Facebook chief Mark Zuckerberg pressed the company to “answer these serious allegations and hold those responsible to account if there has been political bias in the dissemination of trending news.” Ironically it is Thune that tried to legislate the “fairness doctrine” out of broadcast media in 2007 (an FCC policy that was never made law, and was dropped due to intense industry lobbying in 2011). With social media it seems political fairness in journalism has become Thune’s new touchstone issue.

Thune’s inconsistency reflects a real difference between the political economies of broadcast media and social media, especially concerning the production or curation of the news. Underscoring this are tandem paradoxes that this post will explore.

In the broadcast era, media producers generated revenue by monetizing the only quantifiable value they could—the size of the passive audience as cultivated by entertainment content. Paradoxically this single data factor (an index of the passive audience measured by the Nieslen ratings system) gave the audience a high degree of influence on the content by driving competition among commercial networks that resulted in a consistent improvement in the program content and news production (I realize some might argue with this point, but go with it for the moment).

In the social media era we have shifted from a passive audience platform to an active participatory system where a combination of networking and software cultivates content from users, generating increasingly valuable and granular data to advertisers. Commercial appeals can be targeted to individuals in real time based on their profile and immediate online activity (their posts, views, likes, shares, etc.).

Algorithms are where the real value lies. Algorithms define action. – Peter Sondergaard, SVP Gartner Research

Paradoxically this higher value generation results in less end user influence over the software platform. New features on Facebook are not developed so much to serve the user as they are designed to drive more “frictionless” data generation increasing value for paying customers and shareholders. The trending topics links are collected, for example, to provide more data generation opportunities. If you click on news about Prince’s death investigation, subsequent advertising latches on to that data to serve up highly related products and sponsored posts (more Prince-related miscellany).

So, contrary to Dallas Smythe’s “free lunch” of programming in the broadcast system vying for audience attention with the most popular content (what Smythe called the “consciousness industries”), the new bargain of social media is actually resulting in lower value to the user audience over time. Facebook is free but that doesn’t mean you don’t pay for it. Rather, its currency is your data (what you choose, what you share and where you go).

[Facebook] never bothered to reckon with the basic responsibilities that journalism entails, nor the ethical and epistemological challenges it presents—probably because they’re messy and inconvenient and might get in the way of optimizing engagement – Will Oremus, Slate

What is worrisome about the recent scandal is that it may push more people to prefer that Facebook’s news curation for trending topics shift completely away from human subjectivity toward the perceived objectivity of automated algorithms.

Algorithms hold a special station in the new technological temple because computers have become our favorite idols. – Ian Bogost, The Atlantic

This is becoming an increasingly critical issue as new research is demonstrating that many people’s experience of the Internet is becoming more and more defined by their use of social media platforms. For an growing number of people their entire Internet experience is occurring completely within Facebook.

This constitutes a new form of commercial capture. While the broadcast medium subjugates production of the public good of journalism to its ratings-driven commercial platform, Facebook is going much further by coding features that increasingly wall off its product (users) from the open Internet where it can’t effectively obtain profitable user data, and curating news content designed to generate more data from user engagement.

We may be exchanging the phantom specter of subjective political bias for an even more pervasive commercial bias of a social networking medium where algorithmic objectivity is predicated on the data collection potential in the dissemination of all news content.

Algorithms aren’t magic. They’re built by humans, they’re maintained and updated and overseen by humans, and they’re flawed like humans. Most importantly, they’re built to serve human ambitions, which are inherently subjective. – Will Oremus, Slate

For all the politicians and pundits that decry political bias and clamor for news objectivity, the form of objectivity promised by automation is entirely contingent on the subjective goals of the ones that code the software.

Flaws in data and algorithms can leave…us susceptible to an especially pernicious form of automation bias. – Nicholas Carr, The Glass Cage

This begs the question: When it comes to news, is human ideology or automated greed a better master?