Since 2016, social media sites including Facebook, Twitter, and YouTube have vowed to crack down on misinformation related to elections. Monday, they faced their first big test, when delayed results from the Iowa Democratic caucus gave rise to partisan infighting, rampant misinformation, and conspiracy theories. Unsurprisingly, things didn’t exactly go according to plan. Twitter struggled to contain viral electoral misinformation and unfounded accusations of vote rigging from Trump allies, while Facebook grappled with disinformation.
As scrutiny over the app Iowa Democrats commissioned to report results from caucuses grew, reporters, political influencers, and other popular Twitter users shared misinformation regarding who was behind the software, helping the inaccuracies go viral. Simultaneously, President Trump’s family and allies took to Twitter to share and amplify unfounded allegations about the legitimacy of the electoral process. Trump campaign manager Brad Parscale, digital director Gary Coby, and sons Eric and Donald Trump Jr., accused Democrats of “rigging” the caucus amid news that results would be delayed.
“Mark my words, they are rigging this thing … what a mess,” wrote Eric Trump. “And by ‘Quality Control’ they mean fixing the results to get the candidate the Democrat Overlords in DC want,” said Trump Jr.
Twitter’s election integrity policies prohibit people from using the platform “to manipulate or disrupt elections, including through the distribution of false or misleading information about the electoral process or when or how to vote.” In a section detailing examples of content that violates these rules, Twitter lists “misleading claims about voting procedures or techniques which could dissuade voters from participating in an election,” and “misleading claims that polling places are closed, that polling has ended, or other misleading information relating to votes not being counted,” among many others.
Ian Plunkett, the global director of Twitter’s policy communications team, said the tweets did not violate Twitter’s policies “as they do not suppress voter turnout or mislead people about when, where, or how to vote.” In response to a question about why the tweets—which appeared to cast doubt on the legitimacy of the electoral process—did not run afoul of Twitter’s rules prohibiting the “distribution of false or misleading information about the electoral process,” Plunkett gave the same answer. When asked whether Twitter had considered using its so-called quarantine feature—which is designed to partially restrict access to tweets from political figures that break Twitter’s rules yet are too newsworthy to remove entirely—to posts that make unfounded claims about ongoing elections, Plunkett said there were no plans to do so at the moment.
"Is there creditable evidence of rigged results?" asked Joan Donovan, head of the Technology and Social Change Research Project at Harvard Kennedy’s Shorenstein Center. "No, but every disruption is a potential political opportunity, where conspiracies and conjecture are easily amplified across social media.”
On Facebook, pages run by right-wing groups and media figures propelled false information about Iowa’s voter registration rolls into virality shortly before the caucus even began. On Monday morning, Iowa Secretary of State Paul Pate disputed a “false” report from conservative activist group Judicial Watch which claimed that registration rates in Iowa were suspiciously high, and issued an official statement on the inaccuracies later that afternoon. However, the misleading report still spread quickly across social media.
X content
This content can also be viewed on the site it originates from.
The press release from Judicial Watch was shared over 20,000 times on Facebook, according to WIRED’s analysis of data from CrowdTangle, a social media analytics platform owned by Facebook. An article about Judicial Watch’s claim by The Epoch Times—a conservative media organization with a history of aggressive support for Donald Trump that was recently banned from advertising on Facebook—received over 46,000 shares, per CrowdTangle data.
Facebook at first told news outlets that the Judicial Watch post did not violate its rules regarding election integrity “because it did not make misrepresentations about how to vote or whether a vote would be counted.” It later said a review of the post by third-party fact-checkers had determined that it was inaccurate and that it would be labeled as “false information” when shared on Facebook. In a statement, Facebook said that after the outside fact checkers rate a post as false, "we reduce its distribution and apply warning labels to anyone who sees it, tries to share it, or already has."
The Judicial Watch and Epoch Times posts were also shared on Twitter, which declined to take them down, as the company said they did not violate its electoral integrity policies.
Later Tuesday, Twitter also announced a new policy banning users from sharing manipulated images and videos that are “likely to cause harm.” If the rule is violated, Twitter may delete the media, add a warning alerting people that it’s misleading, or reduce its visibility. The new policy applies to deepfakes, but also less sophisticated technologies. For example, Twitter will consider whether a video has been slowed down, selectively edited, cropped, or if captions were removed. “We don’t want to inadvertently reduce the scope of our approach to only one type of media, or only media created via a specific method,” Del Harvey, Twitter’s vice president of trust and safety, said on a conference call with reporters.
Twitter will use three factors to evaluate if a photo or video violates the policy: whether it has been fabricated or manipulated, whether it was shared in a “deceptive manner,” and whether the information is likely to impact public safety or cause serious harm. If a user unintentionally tweets a fake video and it’s not likely to cause harm, Twitter says it will add a warning label, rather than remove the content. When users click on the warning, Twitter will direct them to reputable news sources that have debunked the information, when available. Twitter will begin adding the new labels on March 5.
Yoel Roth, Twitter’s head of site integrity, told reporters the new policy is largely preventive. He said Twitter has mostly seen pornographic deepfakes, which violate its existing nonconsensual nudity policy. Still, the new rule is designed to apply to a wide range of images and videos, and it’s not clear how well Twitter is equipped to enforce it.
Updated, 2-6-20, 3:25pm ET: Facebook did not respond to a request for comment prior to publication. After this article was published, Facebook responded with a statement.
Updated, 2-4-20, 6:05pm ET: This story has been updated too identify Ian Plunkett and to clarify some of his responses.
Updated, 2-4-20, 4:10pm ET: This story has been updated with information about Twitter's policy on manipulated images.
- Inside SpinLaunch, the space industry’s best kept secret
- The bird “snarge” menacing air travel
- The iPad at 10: A look back at its first decade
- He wanted a unicorn. He got ... a sustainable business
- The future of death tech has no rules—yet
- 👁 The secret history of facial recognition. Plus, the latest news on AI
- 💻 Upgrade your work game with our Gear team’s favorite laptops, keyboards, typing alternatives, and noise-canceling headphones
