100+ internal training manuals and other docs show guidelines for Facebook content moderators on topics like violence, hate speech, terrorism, porn, self-harm
Leaked policies guiding moderators on what content to allow are likely to fuel debate about social media giant's ethics
Context & Ripple Effects
The Guardian's publication of over a hundred internal Facebook training manuals is the first in what becomes a running series of leaks about how the company governs speech at scale: the documents show moderators working from yes/no rules on violence, hate speech, terrorism, porn, and self-harm — categories that resist binary answers.
Facebook's response comes within a day, with an explainer defending how and where it draws the line, and the pattern continues through later leaks of election-specific "PR fire" guidance and reporting on the small engineers-lawyers-PR team that writes the rules. The leak matters because it moves moderation from invisible back-office work into public debate about the company's ethics.
First-order effects
- Facebook's thousands of contract moderators now operate under publicly readable rulesets, so every edge-case call they make can be checked against the manual — and criticized when the two diverge.
Second-order effects
- Rival platforms face the same transparency demand: once one company's rulebook is public, keeping your own secret becomes its own story, pushing competitors toward publishing or at least formalizing guidelines.
Third-order effects
- If the leak-and-response cycle holds, content moderation hardens into a distinct corporate function — dedicated policy teams writing auditable rules rather than ad-hoc moderator judgment — setting up the governance frameworks regulators later target.
The trend: Platform content moderation is shifting from opaque internal judgment to formalized, leak-exposed policy infrastructure that invites public and regulatory scrutiny.