By Kent R. Kroeger (July 30, 2018)
Facebook and Twitter stock prices shed around $120 billion on Friday, largely due to slower-than-expected user growth.
Both social media companies are blaming programs they’ve implemented to address automated bots and fake accounts, problems exposed during the current Trump-Russia collusion investigation.
According to a Pew Research Center study, two-thirds of tweeted links to the web’s most popular sites are generated by automated bots (i.e, not human beings). These Twitter bots are largely built by private corporations and political organizations.
“The (Pew) bot study is a reminder that a large portion of what happens on Twitter is simply machines talking to machines,” says business and technology writer Chris O’Brien.
In addressing congressional concerns about how many people may have been exposed to Russian social media propaganda in 2016, the Trump-Russia investigation has laid bare the inaccurate audience metrics used by Facebook and Twitter. That is a good thing.
In October 2017, Twitter announced it had included users of third-party applications as part of their user base. According to a Twitter management letter to its shareholders, since the fourth quarter of 2014, this method over-counted Twitter users by “an average of 1 million to 2 million users per quarter.” Facebook has faced similar criticism.
Internet audience measurement services such as Quantcast, Verto Analytics, and Nielsen have reported the methodological flaws in social media audience numbers for years but were too fragmented and unstandardized to shift investor sentiments. Hopefully, that has now changed.
Facebook and Twitter shareholders may not agree now, but in the long-run this stock price drop was a necessary correction as it will force all social media networks to tighten up their audience metrics so they better reflect the actual number of human beings using their social network services.
But here is the problem now emerging with the social media communities:
The mainstream media’s bullying surrounding Russiaphobia has opened the door for Facebook and Twitter to discriminate against people based on their attitudes, preferences and opinions. In trying to placate frenzied media critics, Facebook and Twitter have since last year started to block or delete, not just Russia-sourced accounts, but also “far-right” accounts that arguably promulgate “fake news” and “hate speech.” The problem is, the humans at Facebook and Twitter have built deeply-flawed algorithms that force machines to decide what user accounts to suspend or delete.
At one point Facebook’s ‘hate speech’-identifying algorithm determined that the posting of the Declaration of Independent by a Texas newspaper, the Liberty County Vindicator, qualified as ‘hate speech’ and therefore was tagged for deletion. Only after notifying the Vindicator of its offense did Facebook realize it had made a mistake.
And what was is the ‘hate speech’ found in the Declaration of Independence?
In articulating the colonist case for breaking from the rule of the British King, the Founding Fathers wrote:
He has excited domestic Insurrections amongst us, and has endeavoured to bring on the Inhabitants of our Frontiers, the merciless Indian Savages, whose known Rule of Warfare, is an undistinguished Destruction, of all Ages, Sexes and Conditions.
While Facebook can be commended for admitting the absurdity of their targeting algorithms, their admission has only amplified the exposure of other cases where legitimate Facebook or Twitter users are being blocked or suspended. Twitter was reportedly suspending one million Twitter users a dayduring one of its more recent purges.
And who exactly is being suspended? It is hard to tell since the social media companies don’t share their proprietary algorithms with the public. However, there is anecdotal information emerging that suggests social and political conservatives are disproportionately being affected. Am I sure of this? No. But it increasingly sounding plausible.
Here is just one example of shadow banning that I find particularly unsettling:
There is a website — shadowban.eu — that allows Twitter users to determine if their Twitter account has been ‘shadow-banned,’ which is the act of blocking a user or their content from an online community such that it will not be readily apparent to the user that they have been banned. ‘Shadow-banning’ is part of what Twitter calls its healthy conversation project (I can only imagine what the kids at Twitter have dreamed up as constituting unhealthy speech.)
One interesting shadow banning case has been brought to my attention by Andrew Torba, the co-founder of Gab, an advertising-free social network dedicated to preserving freedom of speech and the free flow of information on the internet.
As more and more Republican activists and politicians were reporting that they were being shadow banned on Twitter, Torba decided to enter the congressional candidates in his district into shadowban.eu service. First, he entered California Rep. Ted Lieu’s Twitter account into the shadowban.eu service and this was returned:
The ‘shadow ban’ indicator is labelled: No QFD Ban. The QFD stands for ‘Quality Filter Discrimination.’ Figure 1 is telling us Rep. Lieu has full service rights on Twitter. Now let us check is Republican opponent, Kenneth Wright (@drwright4congr1):
Dr. Wright is shadow-banned on Twitter. Why?, you ask. After a few phone calls, a Google search and a background check on Mr. Wright, I found no obvious reason for why Mr. Wright is shadow banned.
The National Review’s Jack Crowe has identified other Republicans and conservatives that have been shadow banned, including Donald Trump Jr.’s spokesman Andrew Surabian (@Surabees).
Twitter has been quick to end the ‘shadow ban’ on prominent Republicans that have not passed the ‘healthy conversation’ filter. Still, the fact that this could even occur in the first place is chilling.
Torba is not holding back on what he thinks is going on.
“Twitter is actively interfering in the U.S. midterm elections,” he says.
And how is Twitter doing this?
According to Torba, “Twitter creates these algorithms to hide the tweets of users who are blocked by many other users, among other signals. Prominent left-wing activists then share block lists that they add anyone who follows too many of the wrong people.”
By widely distributing and sharing the block lists with other Democratic activists, says Torba, these activists effectively flood Twitter’s shadow banning algorithm with hundred of thousands of Republican and conservative Twitter users who are then shadow banned.
To be fair, Twitter is already addressing this problem. Nonetheless, it demonstrates the opportunity for the social media services (not just Twitter) to discriminate against people they decide engage in ‘unhealthy’ conversations.
Given the political donation patterns of Silicon Valley executives, I would not be surprised if their definition of ‘unhealthy’ speech discriminates against Republicans and conservatives.
Are conservatives just being snowflakes?
Perhaps. But with Republican Florida Rep. Matt Gaetz’ threat to file a Federal Election Commission complaint against Twitter over the shadow banning, it is unlikely Twitter will dismiss this issue as mere Trumpian paranoia.
Regardless of how well Twitter addresses the possible discriminatory effects of its shadow banning algorithm, Twitter and Facebook users across all political inclinations need to be vigilant in holding social media companies accountable for any discriminatory business practices.
This is not a freedom of speech issue, per se, but if Facebook and Twitter want to continue as valued online communities, allowing political bias to influence who can and can’t participate in these communities is unacceptable.
Is it unhinged hyperbole to suggest Facebook and Twitter, in the current political environment, might morph into safe-havens for liberals and Democrats only?
I don’t think so. In fact, Facebook and Twitter are already going down that path. And if they continue, don’t be surprised if their stock prices take more big hits down the road.