How Twitter’s Algorithm Poured Gas on the Bernie-Warren Spat

How Twitter’s Algorithm Poured Gas on the Bernie-Warren Spat

January 15, 2020 Off By David Uberti

Want the best of VICE News straight to your inbox? Sign up here.

An awkward exchange between Sens. Bernie Sanders and Elizabeth Warren on the debate stage Tuesday night sparked the type of digital spat that tech platforms can easily blow out of proportion.

#NeverWarren was the top trending topic on Twitter across America Wednesday morning, earning top placement in users’ mobile apps and desktop sites. But there was a hitch: Many of the largest accounts using the hashtag were trying to cool tensions, according to data collected by Hoaxy, a tool created by Indiana University researchers to track the flow of information online.

The accounts still fed Twitter’s algorithm, gave the impression of a sexist onslaught against Warren, and fueled media coverage to drive it home.

Warren and Sanders’ 2020 presidential campaigns have quibbled in recent days over whether he told her in 2018 that a woman couldn’t win the presidency. Tuesday night’s debate moderators raised the issue again: Warren confirmed the story, and Sanders denied it. And they closed the night with a seemingly tense exchange that ended with Warren not shaking Sanders’ hand.

In the hours afterward, Twitter users shared #NeverWarren in an apparently serious reference to the standoff. Others used the hashtag #WarrenIsASnake or posted snake GIFs in her mentions. A handful of pro-Trump accounts amplified both.

But Hoaxy suggests that many of the largest vectors for #NeverWarren were actually users who denounced it. The tool hooks in to Twitter’s API and maps out how hashtags travel across the platform, visualizing engagement — or communication — among different groups of users. NBC News reporter Ben Collins first noted the dynamic.

With #NeverWarren, the largest nodes of discussion on the map tended toward progressive activists and journalists with large followings, who argued that any divide between Warren and Sanders paled in comparison to a Democratic candidate and President Donald Trump. George Takei, the actor and liberal activist, urged users to “knock it off” in a tweet with nearly 10,000 likes and retweets.

The three below added roughly 40,000 more likes and retweets. Those numbers, which are viral by Twitter standards, don’t include mentions:

The tweets that ensued could reopen old wounds for Sanders supporters, who were dogged in 2016 by the “Bernie Bro” caricature of extremely online misogyny. Disinformation researchers have also warned that such cultural debates are fertile ground for attempts to sow division online.

Despite agreeing on nearly everything else, Warren and Sanders’ spat about their previous conversation about sexism and politics has blown up in recent days. Outlets including The Hill, Newsweek, Breitbart, and The Daily Wire all referenced the #NeverWarren hashtag as evidence that the feud between the two candidates, as per CNN’s words, “just got way uglier.”

These Twitter trends, decided by an algorithm and tailored to individual users, are intended “to help you discover the hottest emerging topics of discussion,” according to the company’s website. A Twitter spokesperson added to VICE News that it hadn’t detected any bot activity pushing the hashtag Wednesday.

“People choose to tweet with a hashtag they might disagree with, and our Trends product neutrally represents their behavior in the form of a trending topic,” the spokesperson said.

In other words, a trending topic connotes discussion around a term, not a particular argument for or against that term. That distinction helped turn #NeverWarren into a circular firing squad.

Cover image: Democratic presidential candidate Sen. Elizabeth Warren, D-Mass., left and Sen. Bernie Sanders, I-Vt. talk Tuesday, Jan. 14, 2020, after a Democratic presidential primary debate hosted by CNN and the Des Moines Register in Des Moines, Iowa. (AP Photo/Patrick Semansky)