Social media platforms can’t stop
mass-shot videos from becoming viral


Following the shooting at the mosque in Christchurch, New Zealand, mosque shooting in the
year 2019, Facebook was widely criticized because it allowed the perpetrator Livestream his
murders for seventeen minutes in total silence. Saturday’s racially motivated
made-for-the-internet mass shooting in Buffalo, New York, went differently.

This time, the perpetrator uploaded his horrific actions to Twitch which is a live streaming video
app that is extremely popular with gamers that was then removed much quicker, less than two
minutes after the incident started as per the firm. The moment Twitch shut down its Livestream,
it had only 22 viewers.

However, it didn’t stop anyone from distributing screens of Twitch live streams as well as the
writer’s notes throughout the web which resulted in millions of views, many of which came from
the links that were shared widely via Facebook as well as Twitter.

“It’s an unfortunate situation because you just need one video to allow this thing to be online
forever and to grow,” said Emerson Brooking who is a senior fellow resident in the Atlantic
Council think tank and a researcher on social media.

It shows that even though the major social media platforms such as Facebook or Twitter have
since Christchurch have become more adept in slowing down the spread of horrifying images of
violence against the masses but they aren’t able to stop it all. Twitch could quickly stop the
shooter’s live video feed due to it being an application made for sharing a particular kind of
content, Live gaming in the first person. Facebook, Twitter, and YouTube are a broader
audience of users and post more diverse posts that are then shared through algorithms that are
designed to increase the spread of viral content. In order for Facebook as well as Twitter to stop
spreading all evidence of this video, it would require that the companies be required to
completely alter the way information is distributed through their platforms.

The unhindered dissemination of murder-related videos on the internet is a crucial issue to
address. For the victims as well as their family members the videos deprive victims of their
dignity during their last moments. However, they also promote the desire for fame among
potential mass murderers, who plot horrendous violence in the hopes of the virality of social
media that promotes their hateful beliefs.

In the past, popular social media sites have become more adept at slowing down and limiting
the growth of these kinds of videos. But they’ve not been completely able to stop the spread,
and probably never can.

The efforts of these companies to date have been focused on identifying videos that contain
violence and then preventing users from sharing the video or editing it. In the instance that
occurred in this particular Buffalo incident, YouTube said it has removed at least more than 400
versions of videos that users have attempted to upload since the afternoon of Saturday.
Facebook has also blocked users from uploading different versions of the video, but they won’t
reveal the number of many. Twitter also stated that it was taking down videos that contain the
same content.

These companies also aid one another identify and then blocking or removing the content of this
kind by sharing notes. They are now sharing “hashes” which are digital fingerprints of a video or
image via their Global Internet Forum to Counter Terrorism GIFCT, also known as GIFCT the
industry group created in the year 2017. If these companies share hashes, they give them the
capability to identify and block violent videos. Similar to how platforms such as YouTube look for
content that is in violation of copyright.

Following the shooting in Christchurch in the year 2019, GIFCT created a new
all-hands-on-deck alarm system, dubbed “content incident protocol. “content incident protocol”
to share hashes in case of an emergency such as an attack on a mass. In the case of the
Buffalo shooting the protocol for content, incidents were set to go into effect at 4:52 pm ET on
Saturday night, two an hour following the shooting had begun. As a result, those who wanted to
disperse the use of the videos attempted to modify the videos to deter the hackers for example
by adding banners or zooming into certain parts of the video companies within the consortium
attempted to counter with new hashes which could be used to flag videos that had been altered.

However, hashing videos can only go only so far. One of the primary methods by which the
Buffalo shooter video made its way to popular social media wasn’t via people who posted the
video in the first place instead of linking to other sites.

In one instance the link to the shooting’s video that was hosted by Streamable which is a less
well-known video website was shared on numerous occasions through Facebook and Twitter
within the first few hours following the shooting. The link received more than 43,000 interactions,
which included shares and likes on Facebook as well as being watched over 3 million times until
Streamable took it down, according to the New York Times.

A spokesperson from the parent company of Streamable, Hopin, did not respond to Recode’s
numerous questions regarding why the platform did not take down the shooter’s footage earlier.
The company did release an announcement that said that such videos do not conform to the
guidelines for community members and conditions of service and that the company is working
“diligently to eliminate them quickly and also to terminate the accounts of users who upload
videos.” Streamable is not an official member of GIFCT.

In a widely shared screenshot of the post, a user shared the fact that they reported a post using
the Streamable URL and an image of the shooting to Facebook within a short time after posting
it but only to receive an apology from Facebook which said that the post did not violate its
guidelines. A spokesperson from Meta said to Recode that posts that included the Streamable
link actually violate the company’s policies. Meta stated that the response to the person who
complained about the link was in error and that the company is now looking into the reasons for

In the end, due to how each of these platforms is created, it’s an endless game of
whack-a-mole. Facebook, Twitter, and YouTube have billions of users. But among those billions
of users, there always will be a proportion of users who seek opportunities to abuse these
platforms. Many social media experts have suggested that the main platforms could do better by
focusing on fringe sites such as 4chan and 8chan which are where the links originated, to detect
and stop the links in advance. Researchers have also suggested the platforms invest more in
their systems to collect user feedback.

Some legislators have blamed the social networks for permitting this video to be uploaded at all.

“[T]here’s an over-feeding on social media sites where hatred breeds more hate and that must
stop,” New York Gov. Kathy Hochul said at a news conference on Sunday. “These media outlets
need to be more attentive in observing the content on social media and the fact that it could be
live-streamed across social media platforms, and not removed within one second suggests to
me that there’s an obligation out there.”

Blocking content and catching it quickly isn’t yet proven feasible. It took Twitch 2 minutes to
bring down the Livestream. That is one of the fastest response times we’ve witnessed thus far
from a social media site that allows people to post live. However, those two minutes were
sufficient for the videos’ links to be shared on more popular platforms such as Facebook or Twitter. The issue, then, isn’t about how fast these videos will be removed and more about how
to avoid the following life they could gain on the major social media platforms.

This is where the basic structure of the platforms rubs with reality. They’re designed to be used
by a large number of people and are ready to be exploited. The time and manner in which this
will change are contingent on whether these companies are prepared to put a wrench into the
machine. It does not appear likely.