One such recording, tracked by the Guardian over the course of Friday, was on Facebook for more than five hours before being taken down, despite being headlined “New Zealand mass shooting”.įacebook said by Monday it had recorded “more than 800 visually distinct videos” and shared them with an industry body, the Global Internet Forum to Counter Terrorism, to let other companies take them down as well.
He said Facebook created a digital fingerprint of the initial livestream, which powered the bulk of the automatic removals and enabled more than 80% of the videos to be blocked before they were publicly posted.īut the technology used to create the fingerprints is fragile and can be defeated by simple methods such as filming the screen and uploading the resulting video. The data underlines the difficulty of keeping such content off social networks, particularly in the immediate aftermath of a global news event.įacebook largely relies on viewers to flag problematic livestreams, and does not employ enough moderation staff to watch every live video as it is being aired.Ĭhris Sonderby, a Facebook vice-president, said that in the case of the Christchurch livestream the first user report came in 29 minutes after the broadcast started and 12 minutes after it ended. “In response we took a number of steps, including automatically rejecting any footage of the violence, temporarily suspending the ability to sort or filter searches by upload date, and making sure searches on this event pulled up results from authoritative news sources.” YouTube said it had tried to keep on top of the unprecedented number of videos uploaded, eventually going so far as to eject human reviewers from the loop in order to let automated systems take down more videos instantly.Ī spokesman told the Guardian: “The volume of related videos uploaded to YouTube in the 24 hours after the attack was unprecedented both in scale and speed – at times as fast as a new upload every second.
0 kommentar(er)
