Ten people were killed this weekend racist attack at the Buffalo Supermarket, New York. The 18-year-old shooter, who has become a proponent of white supremacy, hosted a live broadcast of his attack on Twitch, a video game platform owned by Amazon. Although Twitch removed the video two minutes after the violence began, it was still too late – now horrific footage of the terrorist attack is openly spread on platforms such as Facebook and Twitter, even after companies promised to remove the video.

On Facebook, some users who tagged the video received a notification the content does not violate its rules. The company told TechCrunch that it was a mistake, adding that groups work around the clock to delete video footage as well as links to videos posted on other sites. Facebook has said it is also removing copies of the racist flag and content that praises it.

But when we searched for such a simple term as “buffalo shooting footage” on Facebook, one of the first results showed a 54-second recording from a terrorist’s footage screen. TechCrunch came across the video an hour after it was downloaded and reported it immediately. The video was not deleted just three hours after publication, when it had already been viewed more than a thousand times.

Theoretically, this should not happen. A Facebook spokesman told TechCrunch that he had added several versions of the video, as well as racist arrow writing, to a database of intrusive content that helps the platform identify, delete and block such content. We asked Facebook about this particular incident, but they did not provide further details.

“We’re going to keep learning, improving our processes to ensure that in the future we can more quickly detect and remove infringing content,” said Guy Rosen, Facebook’s vice president of honesty, when asked why the company forces delete copies of videos in a conversation that is not related to them on Tuesday.

Repost streams were also easy to find on Twitter. In fact, when we entered “buffalo video” into the search bar, Twitter offered search queries such as “buffalo video, full video graphics”, “buffalo video leak” and “buffalo video graphics”.

We came across several video attacks that have been circulating on Twitter for over two days. One such video garnered more than 261,000 views when we viewed it Tuesday afternoon.

In April, Twitter took a flight which prohibits individual perpetrators of violent attacks from using Twitter. Under this policy, the platform also reserves the right to remove multimedia related to attacks as well as language from terrorist “manifestos”.

“We are removing videos and media related to the incident. In addition, we may remove tweets that distribute a manifesto or other content created by criminals, ”a Twitter spokesman told TechCrunch. The company called the “hateful and discriminatory” content “harmful to society.”

Twitter also claims that some users are trying to circumvent the removal by downloading modified or manipulated content related to the attack.

In contrast, a video of the tragedy over the weekend was relatively difficult to find on YouTube. The main search terms for Buffalo videos were mostly covered by major news agencies. Using the same search terms we used on Twitter and Facebook, we were able to identify several videos on YouTube with thumbnail thumbnails that were not actually related to the content after clicking. The TikTok TechCrunch identified some messages that directed users to websites where they could watch videos, but found no valid footage in the app in our search.

Twitch, Twitter and Facebook have said they are working with the Global Online Counterterrorism Forum to limit video distribution. Twitch and Discord have also confirmed that they are working with government agencies to investigate the situation. Before the attack, the shooter described in detail his plans in the picture on a private server Discord.

According to documents reviewed by TechCrunch, the shooter from Buffalo decided to broadcast his attack on Twitch because the anti-Semitic shooting at Galle Synagogue in 2019 continued on Twitch. more than 30 minutes before it was removed. The shooter reviewed the broadcast on Facebook, but decided not to use the platform because he believed users needed to log in to watch live broadcasts.

Facebook also unintentionally carried out mass shootings that evaded algorithmic detection. In the same year when the shooting at the Galle Synagogue, 50 people were killed Islamophobic attack at two mosques in Christchurch, New Zealand, which broadcast 17 minutes. At least three criminals mass shootings, including a Buffalo suspect, cited the Christchurch massacre as a source of inspiration for their racist attacks.

Facebook noted the day after the Christchurch shooting that it deleted 1.5 million video attacks, 1.2 million of which were blocked on upload. Of course, this raised the question of why Facebook was unable to immediately detect 300,000 such videos, which means 20% rejection.

Judging by how easy it was to find a video of Buffalo shooting on Facebook, it seems the platform still has a long way to go.