[HELICOPTER BLADES WHIRRING] [EXPLOSION] This is a war that’s being uploaded on the internet every day. Evidence of crimes against humanity and war crimes. FIfty-eight countries urged the United Nations to refer these war crimes to the International Criminal Court. Content moderation is kind of like the referee of the internet – calling what’s in and out of bounds and removing inappropriate posts. Let’s say I’m on a YouTube binge and I see a video with child porn in it. I report it to YouTube. That’s called flagging. The video then gets sent to content moderators whose job it is to review flagged videos. And they then decide whether to leave it up or remove it from the platform. Well, it got complicated in 2017. That year, YouTube started relying much more on machine-learning technology to flag content. “No amount of people that we can hire will be enough to review all of the content. We now have A.I. systems that can identify and take down 99 percent of the Al Qaeda- and ISIS-related content in our system before someone, a human, even flags it us. I think we need to do more of that.” Now, Facebook, YouTube and Twitter get a huge amount of content uploaded to their sites every day. YouTube gets 300 hours uploaded per minute. These companies are under a ton of pressure from governments to quickly get rid of the harmful stuff. But it’s become a double-edged sword, and essential human rights content is getting caught in the net. The bottom line is, computers may be good at detecting violence but they’re just not as nuanced as humans. They’re not as good at figuring out whether a video is ISIS propaganda or vital human rights documentation. [PENSIVE MUSIC] [PENSIVE MUSIC]