Post No Evil

Episode Summary

Executive Summary: This RadioLab episode examines how Facebook built and continually revised a global content-moderation rulebook to handle speech, nudity, gore, hate speech, and newsworthy violence. Through cases like breastfeeding protests, hate-speech disputes, the Boston Marathon bombing, and Mexican cartel videos, it shows the tension between objective rules, cultural context, and human judgment at internet scale.

Main Topics: Facebook’s evolving rulebook (Priority: 5/5): The episode centers on Facebook’s internal moderation guidelines: a written, constantly amended system meant to standardize what can stay up or be removed across billions of users. Breastfeeding and the origin of policy precision (Priority: 4/5): A 2008 protest over breastfeeding photos being removed as pornographic forced Facebook to define nudity more explicitly, revealing how vague rules become detailed through pressure and edge cases. Content moderation as industrial labor (Priority: 5/5): The story follows moderators in Manila, Dublin, and elsewhere who review massive volumes of flagged content quickly, often under NDAs, and experience psychological harm from graphic material. Hate speech and protected classes (Priority: 5/5): A comedy-driven dispute over phrases like ‘men are scum’ exposes Facebook’s attempt to adapt U.S. anti-discrimination logic into platform rules, including confusing exceptions and modifiers. Newsworthiness vs consistency (Priority: 5/5): The Boston Marathon bombing photo and a Mexican beheading video highlight the conflict between strict rules and exceptions made for public interest, exposing Facebook’s role as a de facto publisher. Global culture vs universal policy (Priority: 4/5): The episode argues that rules built for one legal/cultural context fail elsewhere because notions like privilege, offense, and harm vary widely across countries and communities.

Key Arguments: Facebook’s moderation system functions like a living common law: rules are repeatedly amended as new examples break older definitions. Vague concepts such as nudity, gore, or hate speech can only be operationalized through narrow, image-based or keyword-based criteria that are imperfect by design. The company prioritizes consistency and scalability over perfect justice, because moderators have only seconds to decide and millions of posts to process. Human moderators suffer real psychological effects from exposure to graphic content, even when automation takes over part of the workload. Facebook’s speech policies are shaped by U.S. legal concepts like protected classes, but those concepts become unstable when applied globally. Once Facebook makes newsworthiness exceptions, it begins acting like an editor or publisher, even while insisting it is only a tech platform. No single rule can be fair in every locale; what counts as harm, truth, or dignity differs across cultures and political contexts.

Data Points: Facebook user count in early period: about 10 million users - Described the company’s size when early site-integrity moderation began Facebook user count later in episode: 2.2 billion users - Referenced as the scale of the global rulebook Early moderation team size: 12 people - Initial site integrity team handling flagged content Expanded moderation workforce: around 16,000 people - Size of moderation operations across outsourced teams Rule update frequency: up to 20 times a month - How often Facebook amended policy rules Content review speed: every three or four seconds - Moderator pace while reviewing flagged items all day Daily flagged content volume: around a million flagged bits per day - Collective moderation workload across teams Terrorist propaganda takedown rate: 99% before user flagging - Automation performance on terrorist content 2013 Boston Marathon bombing date: April 15, 2013 - Incident used to test Facebook’s gore/newsworthiness rules Facebook protest membership: over 10,000 members - Online breastfeeding advocacy group coordinating the protest Virtual protest participation: 12,000 members - Members who changed profile images/statuses in support Mod hours in the Philippines: 2 a.m. to 4 a.m. - Marie’s overnight shift reviewing disturbing material Moderator pay: $2.50 per hour - Marie’s reported wage for content moderation work

Pivotal Quotes: "There are no exceptions. There’s either mistakes or improvements." — Facebook employee/source: Used to explain the company’s refusal to treat newsworthiness as a discretionary carve-out within moderation policy "You can say whatever you want about black children. Whereas in the case of white men, you've got gender and race, both protected, you can't attack them." — Facebook policy explanation: Illustrates the logic of protected classes, modifiers, and the controversy over hate-speech enforcement "We’re not trying to be perfect here, and we’re not even necessarily trying to be 100% just or fair. We’re just trying to make something that works." — Facebook source: Summarizes the utilitarian philosophy behind the moderation system

Implications: The episode suggests platform governance is now a form of global lawmaking under extreme time pressure. For users and industry, that means moderation will remain inconsistent, culturally contested, and politically consequential.