Social media has significantly changed the way we communicate and consume content. Sometimes we even get our news from Facebook and Twitter before actual news sites. Unfortunately, it has also allowed not just misinformation to creep in but, worse, harmful content. After receiving a lot of flak for what is perceived to be complacency, Facebook is tightening the noose on users of its Live streaming platform to prevent it from being misused in such manner again.
Live streaming is all the rage on social media these days and its original intent is to let friends and family experience an event even if only remotely. But last March, the new media was used to broadcast a heinous crime in Christchurch, New Zealand. Facebook was severely criticized for taking to slow to respond to the posting of the live feed and in hunting down copies of the video that circulated on its network.
Facebook has announced two measures to address those lapses. First, it will be more strict with who gets to use Facebook Live. It will be implementing a “one strike” policy so that anyone that violates any of the serious Facebook policies, like promoting terrorism, would immediately be banned from Live for a short duration, even if it’s their first offense. That short period starts at 30 days and could even be forever, depending on repeat offenses.
The Christchurch shooting, however, revealed one other flaw in Facebook’s system. Other Facebook users, for whatever intention or without knowledge, re-posted videos of the shooting even after Facebook took the originals down. These videos were specifically modified to escape detection by Facebook’s automated system.
Unfortunately, Facebook doesn’t have an answer for this yet so its second step is to invest $7.5 million in research partnerships with The University of Maryland, Cornell University, and The University of California, Berkeley. These partnerships aim to develop techniques to detect manipulated videos, images, and audio to improve Facebook’s system. Additionally, the techniques will also try to distinguish adversaries from unwitting posters who re-post content without giving a second thought to the consequences.