There are posts circulating, for example, that — without evidence — raise doubts about whether Trump is being honest about his diagnosis or that speculate that the episode is part of a plan for Trump to retreat before declaring war on his political enemies buried deep inside government.
The companies began moving nearly immediately after Trump’s tweet. YouTube started responding to searches for Trump and Covid with authoritative news videos from sources like CBS and the BBC “within minutes of their diagnosis being made public,” spokesperson Ivy Choi said. TikTok focused its so-called elections war room on tracking Trump’s diagnosis within 20 minutes of the tweet, according to spokesperson Jamie Favazza.
And Facebook, the world’s largest social network, stood up an ad hoc “operations center” — like the ones built around presidential debates and primary election contests — to tackle the Trump-Covid news well before the work day started in Washington, spokesperson Andy Stone said.
Some calls are easy, the companies say. Posts hoping President Trump will lose his life to the virus are banned under the major platforms’ restrictions on bullying or hateful conduct.
“Content that wishes, hopes or expresses a desire for death, serious bodily harm or fatal disease against an individual is against our rules,” Twitter said in a statement. Facebook and TikTok have similar policies.
For those that provided clearly false or misleading information, the platforms employed a technique they adopted in the aftermath of the disastrous 2016 presidential election: outsourcing the calling of balls and strikes to third party fact checkers.
On Friday, a widely circulated post said that military “Doomsday planes” were spotted on the east coast of the U.S. — put there to defend against attacks from foreign adversaries taking advantage of a supposedly incapacitated Trump. Facebook attached a “Missing Context” label to it, based on reports from fact checkers including The Associated Press.
Similarly, Instagram, which is owned by Facebook, put a “False Information” label on posts suggesting that “The Simpsons” animated television program once featured a cartoon Trump in a coffin, citing outside fact checkers.
Other content, the companies say, is more challenging, like posts arguing that Trump is using his diagnosis — perhaps falsely — to avoid his Oct. 15 debate with Joe Biden or to gin up sympathy before the November election.
Many such claims fall into a bucket that YouTube and others call “borderline content” — posts that toe right up against the line of what the platforms prohibit but don’t cross it.
Conspiracy theories pose a particular challenge for the platforms for many reasons, including that it can be difficult to disprove a negative and they often are cloaked in coded language that is meaningful only to their targeted audience.
In many cases, the sites respond by attempting to limit the spread of that sort of material, like by ranking it lower in search results, rather than getting rid of it altogether.
The companies are facing tough calls — sometimes where questionable posts are coming from high-profile accounts. Sen. Kelly Loeffler (R-Ga.) tweeted Friday morning, “Remember: China gave this virus to our President @realDonaldTrump and First Lady @FLOTUS. WE MUST HOLD THEM ACCOUNTABLE.”
A Twitter spokesperson said that the tweet didn’t violate any Twitter rules and thus no action would be taken on it.
Asked about another eye-catching post — one circulating that said that Trump’s current quarantine is part of a plan, detailed by the group QAnon to hide away before launching battle against a supposed pedophile ring run by prominent Democrats — Facebook spokesperson Andy Stone said that the post was going through Facebook’s third-party fact-checking process, adding that how long that took was up to the fact checkers.
In the meantime, said Stone, the post’s distribution has been reduced, including it ranking lower in Facebook’s algorithm-driven news feeds as it otherwise might. The major platforms argue that they’ve generally figured out how to push that sort of conspiracy thinking to the margins of their sites.
And some prominent social platforms said that they aren’t seeing any meaningful misinformation or questionable content pop up around the Trump diagnosis.
“We haven’t seen a ton of this content yet thankfully,” said TikTok spokesperson Jamie Favazza, pointing to the generally light vibe of the platform and the lack of traction there for political disinformation.
Many of the social media companies have been upbraided by critics for not taking more seriously the threat of disinformation ahead of the 2016 election, and for being slow to come to terms with threats to this election.
Facebook, for example, was widely criticized for taking months before deciding to add flags calling into question some of Trump’s posts that sowed doubt about mail-in voting and only moving in August to crack down on QAnon groups on the platform.
Graham Brookie, the director of the Atlantic Council’s Digital Forensic Research Lab, said it’s too soon to judge how well the major platforms performed in the first 24 hours of the Trump diagnosis news because it typically takes days for disinformation to move from the Internet’s fringes to the mainstream.
But Brookie says the worry is the platforms retreat to their old playbook for handling bad political information. Said Brookie, “I would hope that their approach to public health misinformation is the standard that is applied, as opposed to their approach to political content,” which he called “reactive” and “backfooted.”
But one big advantage they had this time around: time zones. The companies said they were helped by the fact that they were getting the news of Trump’s tweet — sent around 10 p.m. California time — when much of the rest of the country was asleep. That allowed them to get a jump on preparations before much of the United States woke up.
The platforms are so massive and content moves so quickly, however, that the deluge can make getting ahead of misinformation a monumental task. Trump’s tweet announcing his diagnosis — his most popular tweet ever — was retweeted more that 894,000 times.
Mark Scott contributed to this report.