Three Big Discussions We Need to Have ASAP About AI and Social Media Moderation - PEN America

What does that mean during the pandemic? AI-reliant moderation being “turned on” globally means that the further you get away from Silicon Valley, there’s a higher likelihood the quality of the content moderation would have gotten worse. While data is scarce on these points, it’s hard to escape the question of how much of the world is currently effectively unmoderated on Facebook, Twitter, and other platforms. That means the potential of unchecked hate speech, overmoderation of certain communities and dissenting voices, and increased vulnerability to disinformation campaigns. Data has not been forthcoming on this from the social media companies, but we know that the Secretary General of the U.N. has cited an unchecked “tsunami of hate” globally, in particular online. The silencing of free expression by those targeted should alarm us all. As with Issue 1, it’s not necessarily that we simply want a return to the pre-pandemic way of doing things—those were bad, too. But some transparency, access to relevant data for researchers and the public, and a good honest conversation could do a world of good.