The US Senate is concerned about political bias at Facebook. Specifically, it wants to know how the company chooses news stories that appear in its “trending topics” section.
The Senate is right to be concerned. Media owners have long held sway over the political process. William Randolph Hearst’s newspapers commanded 20 million readers a day (paywall) at their peak in the 1930s, nearly a fifth of the population. Mark Zuckerberg’s reach is even larger: Facebook had 173 million daily active users at the end of March in the US and Canada, half of those countries’ combined populations.
But whereas Hearst loudly and unabashedly transmitted his political views through his media properties, Zuckerberg’s Facebook is at pains to paint itself as a neutral platform. In defending itself against the Senate’s claims of bias, it said it was ”technically not feasible” for human curators to introduce any. Machines, in other words, have no bias.
The Senate seems to agree with this premise, seeking details on how the human curators do their jobs. The politicians are looking for a human target, a Hearst, to attribute bias to. But in doing so, they miss the point of how Facebook dominates our media diets.
“Trending topics,” after all, is a small slice of what people consume on Facebook. The main meal is its news feed, whose contents are decided almost entirely by machine, a fact to which most Facebook users (pdf) are oblivious.
Technology historian Melvin Kranzberg argued that technology is neither good nor bad, but nor is it neutral; it has consequences that can go far beyond those intended or foreseen. Even Facebook’s creator may not even fully understand the bias, if any, expressed by his invention.
Given Facebook’s power, any scrutiny of it is welcome. But it’s the algorithm that has to be held accountable, not the ghost of 20th-century media barons.
This was published as part of the Quartz Weekend Brief. Sign up for our newsletters here, tailored for morning delivery in Asia, Europe & Africa, and the Americas.