Allen, Texas mall shooting: How graphic videos spread on Twitter

Mourners pray at a memorial to for the victims of the shooting at a mall in Allen, Texas

Graphic videos from the scene of a mass shooting in Texas spread rapidly and were viewed millions of times on Twitter before the social media site began taking the footage down more than 24 hours after the attack.

One video showed dead bodies of the victims, including what appeared to be children. Another showed the perpetrator lying dead on the ground.

Both clips were visible at or near the top of search results in the hours after news broke of the attack north of Dallas.

Many users expressed shock at how easily the footage spread, mostly without content warnings.

"I'm currently feeling overwhelmed by the horrific videos circulating around Allen, Texas," one wrote. "Twitter's moderation team needs to step up."

"I've also been noticing a lot more shocking and sometimes gory content on Twitter showing up on my feed," said Amarnath Amarasingam, an extremism expert at Queen's University in Toronto.

Staff cuts at Twitter, he suggested, meant the company no longer had the expertise to advise on nuanced content removal or when to add warning messages on sensitive and graphic content.

Twitter's chief executive Elon Musk says he has cut more than 6,000 jobs, around 80% of the workforce.

The accounts spreading the video clips had a variety of motivations. One viral tweet - viewed at least 3.6 million times according to Twitter's metrics - came from a pro-Trump journalist who wrote "Something needs to change".

Another came from a Democratic Party activist who wrote: "Seeing a video of slaughtered innocents — including a child — is truly horrifying… But maybe — just maybe — people NEED to see this video, so they'll pressure their elected officials until they TAKE ACTION."

He called for tighter gun control restrictions in a post that was viewed three million times.

• Officials looking into Texas mall gunman's ideology

• Indian engineer among Texas mall shooting victims

• Twitter insiders: We can't protect users from trolling under Musk

The graphic videos were also used to spread false information and boost the popularity of fringe accounts. In one instance, a post from an account with around 1,000 followers was viewed at least 1.3 million times before it was deleted by its author.

It falsely claimed that the assailant was a black supremacist who shouted slogans against white people before the shooting. A similar tweet from a parody account got 70,000 views despite only having 100 followers.

In reality, investigators are looking into whether the killer had neo-Nazi or white supremacist links. The suspect, Mauricio Garcia, aged 33, wore a patch that said "RWDS", short for "Right Wing Death Squad", a slogan used by many far-right activists.

Misinformation expert Marc Owen Jones, associate professor at Hamad Bin Khalifa University in Qatar, says that similar false messages were spread by a number of accounts, including some that had paid for the company's premium "blue tick".

"I don't necessarily think there is something different under Musk with regards to violent content," he said, "but I was also surprised that some of the graphic content lingered after being so heavily retweeted." This should be low-hanging fruit for moderators, he added.

Twitter is a member of Christchurch Call, an organisation founded by the New Zealand and French governments after the 2019 Christchurch shootings with the aim of preventing terrorist and violent extremist content from spreading online.

The tech companies signed up to a pledge to review algorithms and other processes "that may drive users towards and/or amplify" such content.

The company no longer has a press team that responds to questions, and neither Mr Musk or the company's communications account responded to a tweet.

Late on Sunday, Twitter moderators approved at least one viral copy of one of the graphic videos, stating in response to a user report that it did not violate the site's policy on sensitive content.

By Monday, the videos containing the most graphic shots of dead bodies did start disappearing from the platform.

Instead, a new trend was dominating search results for the shooting: edited clips showing a few brief frames of graphic material, with messages urging users to click on links to view the rest. The links led mostly to spam websites.