How the Anti-Defamation League Is Fighting Election Misinformation Online

If you’ve been worried about disinformation or even violence threatening the legitimacy of this presidential election, you’re not alone. Hate speech watchers like the Anti-Defamation League are on high alert.

As several critical battleground states continue to count ballots, President Trump and some of his supporters continue to circulate misinformation about the election — lies which resulted in Twitter applying warning labels to the president's tweets and permanently suspending an account belonging to former White House chief strategist Steve Bannon. Facebook decided to remove the #StoptheSteal campaign altogether.

But the lies are not just burning in the crispy underbrush of Twitter and Facebook.

"[They're also moving] on TikTok, Parler and Telegram. Think of how many platforms you are aware of. Multiply that by 10," said Oren Segal, vice president of ADL's Center on Extremism.

Take, for example, Twitch, best known as a streaming site for video gamers and the people who like to watch them play and talk. Segal reports some of that talk during election season has been political. When that political talk veers into hate, harassment and misinformation, ADL representatives say they reach out to the Amazon-owned company's trust and safety team to make sure they're aware of what's happening and taking steps to address it.

"Where do you go when you want to understand where the threats of extremism are coming from? There's 50 to 100 answers, but they're all online platforms," Segal said.

Are the bigger platforms we're all familiar with enforcing the election season policies they’ve put into place? To varying degrees of effectiveness.

Dave Sifry, vice president of the ADL's Center for Technology and Society, said that some YouTube channels — including ones with more than 1 million subscribers — live-streamed fake election results to tens of thousands of people on Election Day.

"Some of those channels were even able to make money from ads," he said.

Sifry took issue with another YouTube election season policy regarding the labeling of potential misinformation.

"Even when they take the ads off, [the label] says things like 'Results may not be final. See the latest on Google,' " he said, steering viewers to the search engine owned by YouTube's parent company. That's in contrast with other platforms that link directly to election results, or company statements about misleading and inaccurate information.

A YouTube spokeswoman responded that the platform takes down livestreams that violate its community guidelines, in accordance with established policies prohibiting spam, deceptive practices and scams.

“All search results and videos about this election are accompanied by an information panel noting that election results may not be final and linking to Google’s election results feature. The language is similar to what other companies are using," wrote YouTube's Ivy Choi. "The vast majority of elections-related queries are surfacing authoritative sources and the majority of time, borderline content and harmful misinformation are not widely recommended by our systems."

But what happens when a company fails to address flagged misinformation? Or fails to address it comprehensively?

Since the polls opened on Tuesday, Twitter has seen a flurry conspiracy-charged hashtags related to mail-in ballots.

"Now, while we definitely saw Twitter taking action, labeling misleading or false information relating to the election on various prominent accounts, we found 152 tweets for which no action had been taken," Sifry said. "I'm not talking about small accounts that have few followers. We're talking about 84 different, high-visibility accounts with a combined total of 118.8 million followers."

Around the time the ADL initially alerted Twitter to the tweets, they'd received nearly 2 million engagements (retweets, likes, etc.)

"Upon rechecking [the following morning], 67 of those 152 tweets had been labeled as misinformation or annotated with election security information. But the 85 tweets to which Twitter did not respond received an additional 350,000 engagements," Sifry said.

Twitter spokesman Trenton Kennedy responded, "Our teams are reviewing all the Tweets referenced ... and will take enforcement action if they find violative content, in line with the Twitter Rules and our Civic Integrity Policy."

As for misinformation spreading via search engines, ADL Senior Vice President of Policy Eileen Hershenov said the organization is tracking keywords related to things like violent militia groups and known conspiracy theories.

"The top 10 states where we're seeing this: Arizona, Texas, Nevada, Idaho, Washington, Wisconsin, Arkansas, California, Oregon and West Virginia," Hershenov said

Hershenov added that individuals who spot troubling conversations can either go directly to law enforcement agencies and social media platforms like the ADL does, or direct their concerns to the ADL as a middle man.

"There are a lot of people blowing off steam, right?" Sifry said. "That's totally OK. So we are in some sense in the business of finding needles in haystacks."

And what happens after we do know the final outcome of the election? For the White House and other races, too?

"You know, this is not a time to let our guard down," Segal said, warning that people inclined to concoct conspiracy theories will keep at it, long after the rest of us have moved on with our lives.

Especially known conspiracy theorists just elected to Congress, Hershenov added. The ADL, she said, sent a letter to congressional leaders on the issue.

"Look, we have free speech. They can say what they want. They can the assemble and associate with who they want. But we should not give them a platform in Congress. You should not put them in leadership positions, in committees, in caucuses," Hershenov said.