The Technology 202: Twitter’s response to Trump’s coronavirus diagnosis highlights inconsistencies in company’s handling of abuse

From Democratic Reps. Rashida Tlaib (Mich.) and Alexandria Ocasio-Cortez (N.Y.):

Reps. Ilhan Omar (D-Minn.) and Ayanna Pressley (D-Mass.) also responded to the company’s tweet.

Twitter on Saturday said it would do better:

The episode highlighted the broader issue of social media abuse directed at female politicians — particularly from minority backgrounds. 

Female congresswomen are far more likely than their male counterparts to be targeted with abusive posts on Facebook and Twitter, according to a new analysis from the Institute for Strategic Dialogue shared exclusively with The Technology 202. And the research shows that Ocasio-Cortez and Omar received the highest proportion of abusive comments. 

The findings are particularly important in the final weeks of a contentious presidential election, where a Black and Asian American woman is for the first time on the presidential ticket. They’re also a reminder that vice presidential candidate and Sen. Kamala D. Harris (D-Calif.) is particularly vulnerable to online vitriol, especially as the spotlight intensifies this week because of the vice presidential debate. 

The report’s authors note their findings have stark implications in light of other research revealing online abuse can cause lawmakers to step down or otherwise limit their political activities. 

The report’s authors analyzed all tweets specifically tagging ten members of Congress during an 11-day period between June 25 and July 6. They also used the Facebook data aggregator CrowdTangle to review content from public pages and Facebook groups during the same time period.

• Abusive messages accounted for more than 15 percent of those directed at every female lawmaker analyzed, compared with around 5 to 10 percent of male candidates. The only exception: 27 percent of the messages received by Senate Majority Leader Mitch McConnell (R-Ky.) were classified as abusive.
• Women of ethnic minority backgrounds were particularly likely to be targeted. Omar received the highest proportion, 39 percent, of abusive messages of all the candidates. Ocasio-Cortez received the highest ratio of abusive comments on Facebook.
• Male politicians of ethnic minority backgrounds weren’t more vulnerable than White counterparts. Sens. Cory Booker (D-N.J.) and Tim Scott (R-S.C.), who are both Black, received similar levels of abuse to White male candidates. However, the attacks against them were more likely to be about race, the report’s authors found.
• Abuse directed toward women was more likely to be about gender than that targeting men. Abuse targeting men was more generalized and focused on their political stances, while the messages directed at women were more likely to focus on appearance or general competence.
• Female Democrats received ten times more abusive comments than their male counterparts on Facebook. And Republican women received twice as many abusive comments as Republican men.
• The researchers detected more abuse on Twitter than Facebook, but they noted that might be due to the different ways the companies allow researchers to access data.

“The high level of abusive content directed at public figures who are in the political limelight demonstrates not only the fraught, polarised political context in which the election is taking place, but also the platform that social media companies provide to those seeking to attack political candidates and often to threaten their safety,” the report’s authors wrote. “The ability to target and amplify hate and abuse towards political candidates has been hyper-charged by the business models of social media companies that grow quickly, promote outrage or sensationalism, and are unable to deal with the subsequent harms rife across their platforms.”

• Companies should provide greater transparency about their content moderation policies. Specifically, they should give examples of content that falls inside and outside the current rules.
• The companies should archive and preserve all data related to content takedowns during the coronavirus pandemic through artificial intelligence so that people can study the efficacy of these programs.
• They should expand training for human moderators on reviewing abuse toward high-profile individuals.
• The platforms should implement specific measures to reduce harassment of politicians.

The report’s authors also say democratic governments have a role to play, and they encourage regulation that forces greater transparency of the processes and systems to moderate content. They also call for mandatory government access to algorithmic systems. 

“Abuse of women on the Internet is a serious problem, one we tackle in a variety of ways — through technology that identifies and removes potentially abusive content before it happens, by enforcing strict policies, and by talking with experts to ensure we stay ahead of new tactics,” Cindy Southworth, Facebook head of women’s safety said in a statement. “We know this is a particular challenge and we will continue working to find new solutions.” 

And Twitter said it’s committed to respond to concerns “through action, not empty words.” The company has taken recent steps to limit abuse on the platform, such as allowing people to restrict who replies to their tweets.

“Twitter being abused to instill fear, to silence your voice, or to undermine individual safety, is unacceptable,” Twitter spokesman Trenton Kennedy said in a statement. “We appreciate the open and direct feedback we’re getting, and we will respond to it through further action.” 

Nearly a dozen conspiracy theories or fake posts connected to Trump’s coronavirus diagnosis have gone viral on social media since Friday, Jane Lytvynenko at BuzzFeed News found. The fake and misleading images include a video edited to make it appear as though Joe Biden had covid-19 before the debate. (The former vice president has tested negative multiple times).   

To respond to the onslaught of misinformation, tech companies including Facebook, YouTube and Twitter launched command centers similar to the war rooms they use to address election misinformation, Nancy Scola at Politico reports. 

In many instances, the companies took steps to make misleading content harder for users to find, rather than removing it entirely. Facebook labeled a series of viral claims that Trump had launched “doomsday planes” after his diagnosis as “missing context.” (U.S. Strategic Command confirmed the planes were a part of an existing mission and unrelated.)

The diagnosis also fueled posts by followers of the QAnon conspiracy theory, which falsely suggests that Trump is fighting a deep state of Democrats and celebrities who commit heinous crimes. Many of the posts suggest that Trump’s diagnosis date holds a secret message and that the disease is fake, EJ Dickinson at Rolling Stone reports.

Trump’s diagnosis sits at the convergence of coronavirus misinformation and political misinformation that platforms have struggled to police, experts told The Technology 202.

And so far, the White House has provided little clarity about the president’s health, making it even harder for users to sort fact from fiction, as NBC News’s Ben Collins points out:

Breaking up the company, which includes Instagram and WhatsApp, could cost billions of dollars, according to a 14-page legal treatise company lawyers have prepared in the social media giant’s defense, Jeff Horwitz at The Wall Street Journal reports. Facebook argues unwinding its acquisitions of Instagram and WhatsApp would weaken security and harm users. 

“A ‘breakup’ of Facebook is thus a complete nonstarter,” the company states.

The company also points to the Federal Trade Commission’s decision to close reviews of its 2012 Instagram and 2014 WhatsApp acquisitions without opposition. Unlike Congress, the FTC had access to many of the emails that lawmakers have pointed to as evidence of Facebook’s push to crush its competition.

But Facebook’s critics say the argument is unlikely to hold up under legal scrutiny.

“There’s no way a decision on one merger would be preclusive,” said Tim Wu, a Columbia University law professor who has called for the break up of Facebook. “There is no ‘it’s too hard’ defense.”

The House Judiciary antitrust subcommittee is expected to release the results of its year-long investigation into Facebook and other tech giants this week.

More than 93,000 active accounts reference the conspiracy theory in their Twitter profiles, the nonpartisan research group Advance Democracy found. Researchers also found while the use of traditional hashtags associated with the movement has declined, alternative spellings and variations to avoid detection have increased, Craig Timberg reports. 

Advance Democracy found alternative hashtags such as #CueAnon have surged since Twitter’s crackdown. But they’re still reaching thousands of fewer users than traditional hashtags did before. QAnon supporters have also started to encourage followers to move to less moderated apps, Parler and Gab. 

Twitter reports that its July decision to ban 7,000 accounts and restrict over 100,000 others has cut discussion of the conspiracy theory by half.

“We aim to be iterative and transparent in our approach, and we recognize we cannot review every tweet containing a falsehood,” said Twitter spokeswoman Lauren Alexander. “That’s why, under this expansive framework, we noted that we will prioritize the removal of content that has the greatest potential for harm.”

But social media companies face an uphill battle now that the conspiracy theory has made its way offline into mainstream political discourse. On Friday, all but 17 House Republicans voted to condemn the conspiracy theory

“Addressing this threat is going to require more robust action by the social media platforms, but more importantly, it’s going to require that those elected officials sitting silently on the sidelines stand up and address this threat to our democracy,” said Daniel J. Jones, president of Advance Democracy.

Trump ordered a ban on the popular Chinese messaging and e-commerce app on the grounds that it poses a national security threat.  The White House alleges the app’s Chinese owner Tencent could be compelled to share U.S. user data with the Chinese government.  

A federal judge granted an injunction on the ban last month, siding with WeChat users who sued the government saying the ban violated free speech.

Tencent disputes the White House’s claims.

• The Brookings Institution will hold an event on what to expect on tech policy in the next presidential administration on Tuesday at 2 p.m.

SNL takes on the TikTok ban.