There’s a persistent belief that social media platforms secretly maintain a list of banned words—terms you should never say unless you want your post to disappear into the algorithmic void.
That belief fuels a strange trend you’ve probably seen: creators avoiding ordinary words and replacing them with awkward euphemisms. “Unalived” instead of “killed.” “Pew pews” for guns. “Seggs” for sex. People admit these substitutions sound silly, but many feel they have no real choice.
This coded vocabulary—known as algospeak—is built on the assumption that platforms suppress content containing certain words to push political agendas or present a sanitized environment for advertisers.
The tech companies flatly deny it.
“YouTube does not have a list of banned or restricted words,” spokesperson Boot Bullwinkle told the BBC. “Context matters. The diversity of voices on YouTube shows that a nuanced approach works.” Meta and TikTok insist the same: no secret list, no forbidden phrases, no conspiracy.
But reality is more complicated.
The blurry line between moderation and manipulation
Social media history is filled with examples of platforms quietly influencing which topics rise and which sink—sometimes contradicting their public claims about fairness and transparency. Even if they aren’t targeting isolated words, experts say these companies absolutely intervene in subtle ways.
The biggest problem is the ambiguity.
When a post flops, you can never tell why. Was it poor content? Bad timing? Or did you accidentally trip some invisible rule?
That uncertainty drives widespread self-censorship. At best, people resort to goofy wording to discuss serious issues. At worst, creators avoid entire topics because they fear suppression.
In a world where social media is the primary news source for millions, that silence may mean some ideas never reach certain audiences at all.
One creator’s struggle with the algorithm
Alex Pearlman—who has millions of followers across TikTok, Instagram, and YouTube—says algorithmic censorship shapes his work constantly.
“On TikTok alone, I almost never say the word ‘YouTube,’” he explains. Analytics convinced him that telling viewers to visit a competitor’s platform makes his TikTok videos tank. Many creators believe the same, though TikTok denies it.
But the more disturbing experiences involve sensitive topics. Pearlman frequently posts about Jeffrey Epstein. Last August, several Epstein-related videos were suddenly removed from TikTok in a single day. They stayed up on Instagram and YouTube. TikTok offered no explanation, no highlighted violation—just strikes on his account and rejected appeals.
Soon, Pearlman noticed fewer big creators talking about Epstein at all.
To continue covering the topic, he began using a code name: “the Island Man.”
But, as he notes, coded language often confuses audiences and dilutes the message.
TikTok declined to comment on his specific case, repeating only that such concerns reflect misconceptions about how the platform works.
A pattern of opaque decisions
Despite public assurances of fairness, platforms have repeatedly been caught manipulating visibility.
- BBC and Human Rights Watch investigations found that Instagram and Facebook systematically limited Palestinian content after the 7 October Hamas attacks. Meta called these “mistakes,” not intentional suppression.
- Leaked TikTok documents (2019) revealed moderators were told to downrank content from “ugly,” poor, disabled, or LGBTQ+ users to keep the app looking more “appealing.” TikTok later claimed this was a misguided anti-bullying effort and said the policy was discontinued.
- The same leaks showed TikTok restricted livestreams criticizing governments—though it insisted such rules didn’t apply to the U.S.
- In 2023, TikTok admitted to a secret internal “heating” button that could artificially make chosen videos go viral. If there’s a “heater,” creators argue, there is likely a “cooler,” too.
YouTube has faced similar accusations. LGBTQ+ creators sued the platform in 2019 over alleged demonetization tied to words like “gay” and “trans.” The lawsuit was dismissed, and YouTube denies ever restricting LGBTQ+ content.
When users start fooling themselves
Platforms do intervene—but the rules are vague, shifting, and often invisible. That opacity fuels speculation, folk theories, and increasingly elaborate user responses.
Professor Sarah T. Roberts of UCLA calls social media a space where rules “don’t make sense to ordinary people,” creating bizarre behaviors that only fit within the twisted logic of platform culture.
One striking example happened in August 2025, when thousands of users started posting excitedly about a huge music festival in Los Angeles. They described performances from stars like Sabrina Carpenter, dazzling lights, and massive crowds.
But there was no festival.
The posts were code. That month, protests erupted across the U.S. over ICE raids, and many users—assuming censorship—referenced the demonstrations as a “music festival” to “trick” the algorithm.
A TikTok video by creator Johnny Palmadessa even showed him standing in the middle of a protest saying, “We gotta call it a ‘music festival’ so the algorithm shows you this.”
Ironically, linguist Adam Aleksic noted that TikTok does allow protest coverage. The “music festival” meme mostly arose because people were over-correcting, unsure of what might be censored. And the coded phrase actually boosted engagement, convincing users the censorship was real.
Researchers call this effect the algorithmic imaginary: people change their behavior based on what they think algorithms do—whether or not it’s accurate. Their behavior then shapes the algorithm in return.
Myth, reality, and a lot of confusion
Plenty of content about Epstein, Gaza, and other sensitive topics remains on major platforms. And if TikTok were truly suppressing discussions of murder, the term “unalive” would likely be banned too.
Creators admit they’re guessing.
“We’re just throwing everything at the wall to see what sticks,” says activist Ariana Jasmine Afshar.
Meta did openly suppress political content from 2023 to early 2025 before reversing course after President Trump’s second inauguration. During that period, coded language might indeed have slipped past certain filters.
Afshar says she’s personally seen content suppressed—and yet, paradoxically, Instagram also reached out to congratulate her on her work and offer guidance on how to succeed on the platform.
“It’s real,” she says. “But also confusing. Very confusing.”
So what’s really going on?
According to Roberts, understanding algospeak and algorithmic paranoia requires understanding the platforms themselves.
And ultimately, she argues, it’s not about ideology.
It’s about profit.
Platforms don’t primarily moderate to push political agendas—they moderate to protect their revenue streams, maintain user engagement, and keep advertisers happy. Anything that threatens those goals—controversy, conflict, or discomfort—risks being quietly pushed out of sight.
In that environment, users fill in the gaps with theories, code words, and rituals designed to appease the opaque systems controlling their reach.
And even if the “banned words list” doesn’t exist, millions behave as if it does—reshaping online language, culture, and public discourse along the way.
