Discussion: (0 comments)
There are no comments available.
A public policy blog from AEI
More options: Share,
Sen. Ted Cruz led off Monday’s congressional hearing on Big Tech’s supposed anti-conservative bias with a humble concession: “I will note much of the argument in this topic is anecdotal. It’s based on one example or another example.” And the reason for that, Cruz continued, is insufficient transparency by Google, Facebook, and Twitter. “Nobody knows what the raw data is in terms of bias,” he added.
But that’s not quite right. There’s actually a significant amount of data available to help analyze the bias issue. A Twitter executive at the hearing noted that in 2018 there were 33 million MAGA tweets with #MAGA the fifth most tweeted hashtag. And a recent internal Twitter study found that tweets by Democratic and Republican members of Congress perform pretty much the same.
Other research has determined that partisan Facebook pages have roughly equal engagement, and right-leaning pages in total have a bigger presence on Facebook than those that lean left. Then there’s a Columbia University analysis that looked at “prominent, politically active” people suspended from Twitter since the company’s launch in 2006 — which is probably the sort of info Cruz wants — who meet two other qualifications: they expressed a Trump or Clinton preference in the 2016 presidential election and their suspension was covered by mainstream news sites. But while the study found that 21 of 22 accounts were pro-Trump, they were also a rogues’ gallery of “outspoken or accused white nationalists, neo-Confederates, holocaust deniers, conspiracy peddlers, professional trolls, and other alt-right or fringe personalities,” according to TechDirt.
Indeed, even the anecdotes of bias mentioned in the hearing generally seem to have plausible explanations. Was the Twitter account for the pro-life film “Unplanned” temporarily suspended because of its subject matter or, as Twitter explains, because the algorithm noticed the Unplanned account was registered by someone previously suspended for breaking Twitter’s terms of service rules? As reporter Makena Kelly of The Verge concluded, “… all of the evidence Republicans provided were anecdotal stories that representatives from the companies were able to clearly explain away by citing their respective content policies or investigations into specific takedowns.” Similarly, expert panelist Francesca Tripodi of James Madison University offered this explanation for some of the bias anecdotes, based on her research:
Accusations that Big Tech companies silence conservative focus on output, the idea that Google does not return specific kinds of ideological content. My research demonstrates that Google is not censoring content based on ideology. Instead, I find that the results people get are more likely to be shaped by input — the key words we put into the search bar. … While there are anecdotes that some people or organizations have seen their content removed or minimized — these accusations lack systematic evidence that this was based on some political decision by executives at the top. The content that is produced by conservatives and progressives is different and so are the queries that people espouse those ideals are in searching for.
One complaint by committee Republicans does have validity. Social media platforms are going to have to be more transparent with how posts are flagged and how removal decisions are made, particularly when humans with biases — unconscious or not — are making the call. (Making the algos more transparent does have the risk of making them easier to game or manipulate.) The issue is whether any bias for this or that side basically nets out, how prevalent any bias is, and whether there is any significant harm being caused “from one example or another example.” (That, especially given the potential downside from making platforms more legally at risk from their users’ speech.) And so far, at least, it looks like significant, systemic bias against conservatives isn’t much of problem.
There are no comments available.
1789 Massachusetts Avenue, NW, Washington, DC 20036
© 2019 American Enterprise Institute