Leaked docs: TikTok told moderators to suppress posts from those deemed “ugly”, poor, or disabled, and censor political speech in livestreams, to attract users
The makers of TikTok, the Chinese video-sharing app with hundreds of millions of users around the world … Tweets: @theintercept , @intothecrevasse , @cindygallop , @sarahemclaugh , @rose_catb , and @smearle_rh See also Mediagazer Tweets: @theintercept : Leaked documents reveal the China-based social media company has policies censoring various forms of political speech, and instructed moderators to suppress posts by “ugly” people and the poor to attract new users. https://theintercept.com/... Charles Bramesco / @intothecrevasse : Man, wait until they find out about the rest of the entertainment industry https://twitter.com/... Cindy Gallop / @cindygallop : “Abnormal body shape,” “ugly facial looks,” dwarfism, and “obvious beer belly,” “too many wrinkles,” “eye disorders,” and many other “low quality” traits are all enough to keep uploads out of the algorithmic fire hose.' @samfbiddle @manifesteiro @tatikmd https://theintercept.com/... Sarah McLaughlin / @sarahemclaugh : it's almost impressive how awful this is in so many different ways https://theintercept.com/... https://twitter.com/... Rose Troup Buchanan / @rose_catb : TikTok moderators were explicitly told to suppress uploads from users with flaws both congenital and inevitable. “Abnormal body shape,” “ugly facial looks,” dwarfism, and “obvious beer belly,” “too many wrinkles,” “eye disorders,” and many other “low quality” traits https://twitter.com/... Sam Earle / @smearle_rh : Assuming there's latent demand for an app with a bias toward ugly, poor people and fervent political speech. Somebody make it and cash out. https://twitter.com/... See also Mediagazer Expand More For Next Unexpand More For Next
Context & Ripple Effects
This 2020 leak is the earliest hard documentation of what TikTok's critics later assembled piecemeal: moderation bent to commercial goals rather than stated policy. The near-absence of Hong Kong protest videos on the platform had already suggested politically sensitive topics were being managed; these documents show the same logic applied at the level of individual users' faces, incomes, and livestream speech.
The arc since then has run one direction. A [[a:973646|Beijing engineering document detailing an algorithm optimized purely for retention and time spent]], followed by reporting that executives relaxed moderation rules more than a year before the US election to appeal to Trump supporters, frames today's story as the founding entry in a consistent pattern: enforcement follows whatever grows the audience.
First-order effects
- Users flagged as unattractive, poor, or disabled face direct suppression of their posts under moderator instructions, while political speech in livestreams is censored regardless of its content.
Second-order effects
- TikTok's Washington charm offensive — including transparency-center tours modeled on Huawei's failed playbook — loses credibility every time leaked internal policy contradicts the public-facing version, handing ammunition to the regulators it is courting.
Third-order effects
- If moderation keeps tracking growth and political incentives rather than published rules, platforms drift toward a two-tier system where official policy is marketing and real enforcement lives in internal documents — the structural problem that content-moderation regulation attempts are built around.
The trend: Platform moderation is increasingly revealed as a growth instrument first and a safety system second, with each leak narrowing the gap between what social apps promise regulators and what their internal documents instruct.