Facebook's ineffective content moderation efforts, unevenly enforced and handled by low-paid moderators, show it can't and doesn't have the will to fix itself
The platform is overrun with hate speech and disinformation. Does it actually want to solve the problem? Tweets: @tonyromm , @newyorker , @kevinroose , and @brianstelter Tweets: Tony Romm / @tonyromm : “In retrospect, it seems that [Facebook's] strategy has never been to manage the problem of dangerous content, but rather to manage the public's perception of the problem” https://www.newyorker.com/... @newyorker : “There's no perfect approach to content moderation,” Facebook's former head of content policy said, “but they could at least try to look less transparently craven and incoherent.” https://nyer.cm/BzLOGux Kevin Roose / @kevinroose : Big @andrewmarantz story about how Facebook went from moderating by principle to allowing politicians and P.R. crises to dictate their rules. “Pretty much the only language Facebook understands is public embarrassment.” https://www.newyorker.com/... Brian Stelter / @brianstelter : “Pretty much the only language Facebook understands is public embarrassment.” https://www.newyorker.com/...
Context & Ripple Effects
This piece lands at the end of a three-year arc in which Facebook's moderation machinery kept growing while its credibility kept shrinking. The leaked documents describing its moderation apparatus showed the sheer logistics of screening billions of posts a week in 100+ languages, and reporting on the small internal policy team of engineers, lawyers, and PR staff revealed how few people set rules for everyone else.
What changed is the diagnosis: earlier coverage treated moderation failures as a capacity problem, but the New Yorker — citing Tony Romm's line about managing public perception rather than dangerous content — reframes it as a will problem, reinforced by the research into Facebook's polarizing effect that Zuckerberg reportedly shelved.
First-order effects
- Facebook's own framing is now contested from inside the record: a policy apparatus built on principles has, per the relationships here, been bent around politicians and PR crises, leaving hate speech and disinformation enforcement uneven across the platform.
- The low-paid moderators doing the frontline work face renewed scrutiny over working conditions and quality, since the article ties enforcement gaps directly to how the labor is structured.
Second-order effects
- External parties keep absorbing the shortfall: the related reporting on platforms relying on journalists as unpaid moderators shows the cost of weak internal enforcement being pushed onto newsrooms with none of the resources.
- Rule changes driven by crisis rather than principle — like the later policy letting anyone request takedown of posts showing their residence — become the template, making Facebook's rules reactive artifacts of whatever controversy is live.
Third-order effects
- If the pattern holds, self-governance loses legitimacy as a model: a platform whose documented strategy is perception management invites regulators and outside auditors to take over functions it demonstrably won't perform itself.
- The deeper structural shift is that moderation outcomes stop tracking stated policies and start tracking political exposure — meaning scale alone no longer explains platform behavior, and incentive design does.
The trend: Platform governance is drifting from principle-based internal moderation toward externally imposed accountability, as evidence accumulates that large platforms' incentives favor managing perception over enforcing their own rules.