The price of a "clean" internet | Hans Block and Moritz Riesewieck

Episode Summary

Executive Summary: The talk argues that social platforms rely on hidden human content moderators to police violent, hateful, and ambiguous material that algorithms cannot reliably judge. While this labor protects users, it also exposes workers to severe psychological harm and raises urgent questions about censorship, free expression, and who should govern the digital public sphere.

Main Topics: The hidden labor behind platform safety (Priority: 5/5): Social networks depend on outsourced human moderators to review content that algorithms cannot accurately classify, yet these workers remain largely invisible and silenced. Psychological toll on content moderators (Priority: 5/5): Repeated exposure to graphic violence, abuse, and suicide content can cause trauma, depression, addiction, PTSD, and even suicide among moderators. The limits of automated moderation (Priority: 4/5): Image recognition can detect some obvious features, but it struggles with context, satire, art, activism, and other ambiguous content that requires human judgment. Censorship, safety, and free expression (Priority: 5/5): The speakers question whether deleting disturbing content protects society or erases evidence of injustice and weakens public awareness and debate. Outsourcing moral and political authority (Priority: 4/5): Moderation has shifted from public institutions to private companies and third-party contractors, effectively privatizing decisions about speech and safety. The need for democratic governance of platforms (Priority: 5/5): The talk calls for open, cosmopolitan institutions to set moderation rules rather than leaving these decisions to opaque corporate back rooms.

Key Arguments: Content moderation is essential because algorithms cannot reliably interpret context, meaning humans must make final decisions about what stays online. The people doing this work are deliberately hidden through NDAs, code words, surveillance, and threats, despite performing a critical public function. Moderators are exposed to extreme material at scale, and the psychological damage can be profound and long-lasting. Deleting all disturbing content can also suppress civil rights documentation, citizen journalism, and evidence of injustice. Personalized content filters could fragment shared reality and make collective discussion of social problems harder. The governance of social media should not be left to private companies alone; it requires transparent, democratic institutions. The current model reflects a broader trend of outsourcing responsibility and sanitizing public life by making problems invisible rather than addressing them.

Data Points: Video shares before removal: 16,000 - A rape video on Facebook was shared worldwide before being taken down. Video likes before removal: 4,000 - The same video was liked thousands of times before deletion. Estimated deaths under Duterte since 2016: 20,000 - Used as a comparison for the rhetoric of “cleaning up” society. Moderator training length: 3 to 5 days - The talk says young moderators receive only brief training before handling traumatic content. Moderator age: early 20s - Describes the demographic of many outsourced content moderators. Platform decision options: 2 - Moderators are often reduced to deciding whether content is ignored or deleted.

Pivotal Quotes: "The question of freedom and democracy must not only have these two options. Delete. Or ignore." — Hans Block and Moritz Reisweik: Closing argument about the need for more nuanced governance of online speech. "What Duterte does on the streets, I do for the internet." — A content moderator in the film: Illustrates how moderators see themselves as enforcing order through removal of harmful content. "We will fix that." — Mark Zuckerberg: Referenced as the standard corporate response to criticism of platform moderation.

Implications: Platforms need transparent, democratic moderation systems that protect users without erasing evidence, silencing activism, or harming workers. The future of online speech depends on balancing safety with free expression and shared public reality.