Facebook admits it was 'too slow' to remove murder video and vows review reporting process while it develops AI police
Facebook has admitted that it was “too slow” in taking down a video showing an elderly man being murdered and has vowed to review its reporting process while it develops artificial intelligence measures to better police explicit content.
The shooting in Cleveland is the latest in a string of inappropriate videos to reside on the site
On Sunday (16 April) Steve Stephens shot 74 year-old Robert Godwin at close range in Cleveland before uploading a video of the killing to Facebook. Less than 20 minutes after the video of the murder was uploaded, someone reported a separate, five-minute Facebook Live video of the suspect confessing to the crime.
The video stayed up on the site for almost two hours before Facebook took it down, reigniting the ongoing conversation about the company’s inability to effectively police explicit video content on its platform.
Justin Osofksy, vice president of global operations, confessed that Facebook was “too slow” to act and said, “we know we need to do better”.
In a blog published yesterday (17 April) Osofksy said Facebook is going to be “reviewing our reporting flows to be sure people can report videos and other material that violates our standards as easily and quickly as possible”.
He also brought up the possibility of using artificial intelligence to keep people from sharing graphic videos in the future and to remove videos once they have been flagged.
In February Mark Zuckerberg revealed that the company was working on developing AI which could detect and read video content, however it was still in “very early development”.
Facebook has been reluctant to define itself as a media company and has stated that it does not want to overly arbitrate what is posted on its site, however the reluctance and slowness to intervene with inappropriate content appears to be leaving the door open to people using the platform to show killings and other crimes.