As the pandemic has prompted more people to communicate and express themselves online, algorithmic content modification systems have had an unprecedented impact on the words we choose, particularly on TikTok, and given rise to a new form of Internet-based Aesopian language.
Unlike other mainstream social platforms, the primary way to distribute content on TikTok is through an algorithmically formatted “For You” page; Having followers does not guarantee that people will see your content. This shift has led regular users to essentially customize their videos towards the algorithm, rather than following it, meaning that adhering to the rules of content moderation is more important than ever.
When the pandemic broke out, people on TikTok and other apps started referring to it as “Backstreet Boys Reunion TourOr we call it ‘Panini’ or ‘Panda Express’ as platforms lowered the rankings of videos mentioning the pandemic by name in an effort to combat misinformation. When young people began discussing their struggles with mental health issues, they spoke of ‘not being able to live’ in order to Have Candid Conversations About Suicide Without Account Punishment.Sex workers, who have long been censored by moderation regimes, refer to themselves on TikTok as “accountants” and use the atom symbol as a substitute for the word “porn.”
As discussions of key events are filtered through algorithmic content delivery systems, more users are flexing their language. Recently, while discussing the invasion of Ukraine, people on YouTube and TikTok used sunflower emojis to denote the country. When encouraging fans to follow them elsewhere, users will say “blink in lio” in exchange for “link in bio.”
Euphemisms are especially common in extremist or harmful societies. Pro-anorexia eating disorder communities have long adopted variations of milder words to avoid restrictions. One paper from the School of Interactive Computing, Georgia Institute of Technology found that the complexity of such variables increased over time. Last year, anti-vaccine groups on Facebook began changing their names to “dance party” or “dinner party,” and anti-vaccine influencers on Instagram used similar codewords, referring to vaccinated people as “swimmers.”
Language design to avoid scrutiny predates the Internet. Many religions have avoided pronouncing Satan’s name for fear of being called upon, while people living in oppressive regimes have developed code words for discussing taboo topics.
Early Internet users used alternate spelling or “blank talk” to bypass word filters in chat rooms, photo boards, games, and online forums. But algorithmic content modification systems are more prevalent on the modern internet, and often end up silencing marginalized communities and important debates.
During YouTube’s “adpocalypse” in 2017, when advertisers withdrew their money from the platform over concerns about unsafe content, LGBTQ content creators spoke of having videos demonized for saying the word “gay.” Some started using word less or substituting others to keep their content monetized. Recently, TikTok users have started saying “abundance” rather than “homophobia,” or saying they are members of the “leg booty” community to indicate that they are LGBTQ.
“There’s a line we have to follow, it’s a never-ending battle to say something and try to get the message across without saying it directly,” said Sean Szolek-VanValkenburgh, creator of TikTok with over 1.2 million followers. “It disproportionately affects the LGBTQIA community and the BIPOC community because we’re the ones making that talk and coming up with the seminars.”
Conversations about women’s health, pregnancy, and menstrual cycles on TikTok are also consistently low-ranking, said Catherine Cross, a 23-year-old content creator and founder of Anja Health, a start-up that provides cord blood banking. Replace the words “sex”, “period” and “vagina” with other words or spell them with symbols in the comments. Many users say “nip nops” instead of “nipples”.
“It makes me feel like I need a disclaimer because I feel it makes it look unprofessional to have these oddly written words in your annotations,” she said, “especially for content that’s supposed to be serious and medically inclined.”
Since online algorithms often identify content that mentions certain words, without context, some users avoid pronouncing them altogether, simply because they have alternate meanings. “You have to say ‘saltines’ when you literally talk about biscuits now,” said Loudan Erissian, community manager at Twitch creators. Twitch and other platforms even went so far as to remove some sentiment because people were using it to communicate certain words.
Black and transgender users, and those from other marginalized communities, often use algospeak to discuss the persecution they face, exchanging words for “white” or “racist”. Some get too nervous to say “white” at all and hold their palms toward the camera to point at white people.
said Angel Diaz, a lecturer at UCLA Law School who studies technology and racial discrimination.
In January, Kendra Calhoun, a postdoctoral researcher in linguistic anthropology at the University of California, and Alexia Fawcett, a doctoral student in linguistics at UC Santa Barbara, gave a presentation about language on TikTok. They show how, through self-censoring words in TikToks labels, new algospeak code words have emerged.
TikTok users now use the phrase “le dollar bean” instead of “lesbian” because it’s how TikTok’s text-to-speech feature pronounces “le $bian,” a censored way of writing “lesbian” that users believe will avoid moderating content.
Algorithms cause human language to be redirected around in real time. I’m listening to this YouTuber say things like “The bad guy didn’t revive his followers” because words like “Kill” are associated with demonization
– badidea 🪐 (@0xabad1dea) December 15 2021
Trying to put certain words on the platforms is a fool’s job, said Evan Greer, director of Fight for the Future, a digital rights advocacy nonprofit.
“First of all, it doesn’t actually work,” she said. “People who use platforms to orchestrate real damage are very good at figuring out how to get around these systems. And secondly, it does collateral damage for literal speech.” Greer argues that trying to organize human speech on a scale of billions of people in dozens of different languages and trying to deal with things like humour, irony, local context, and slang can’t be done simply by lowering the rating of some words.
“I feel like this is a good example of why serious moderation is not going to be a real solution to the harms we’re seeing from the business practices of big tech companies,” she said. “You can see how slippery that slope is. Over the years, we have seen more and more misguided demand from the general public for platforms to quickly remove more content regardless of cost.”
The creators of Big TikTok created shared Google Docs with lists of hundreds of words that they believe the app’s moderation systems see as a problem. Other users keep a constant log of terms they believe choked certain videos, in an effort to reverse engineering the system.
“Zuck Got Me For”, a site created by a meme account admin that goes to Ana, is a place where creators can upload meaningless content that has been blocked by Instagram’s moderation algorithms. In a statement about her project, she wrote, “Creative freedom is one of the only silver linings to this burning inferno of the internet, we all are inside…while algorithms stress indie creators who are suffering.”
It also shows how to speak online in a way that avoids the use of filters. “If you violate our terms of service, you may not be able to use vulgar words or negative words like ‘hate’, ‘murder’, ‘ugly’, ‘stupid’, etc,” she said. I often write, ‘I’ the opposite of love xyz” instead of “I hate xyz”.
The Association of Online Creators, a labor advocacy group, has released a list of demands, asking TikTok for more transparency in how it manages content. “People have to skimp on their own language to avoid offending the all-seeing and all-knowing Gods of TikTok,” said Cecilia Gray, creator of TikTok and co-founder of the organization.
TikTok provides an online resource center for creators seeking to learn more about its recommendation systems, and it has opened several transparency and accountability centers where guests can learn how the app’s algorithm works.
In some countries where moderation is heavy, people end up building new dialects to communicate, said Vince Lynch, CEO of IV.AI, an artificial intelligence platform for language understanding. “They become de facto sub-languages,” he said.
But as algospeak becomes more popular and alternative words become popular slang, users are finding that they have to get more creative to evade filters. “It has turned into a mole-hit game,” said Gretchen McCulloch, linguist and author of Because the Internet, a book about how the Internet shapes language. When platforms start noticing people saying “seggs” instead of “sex,” for example, some users have reported that they think even alternate words are flagged.
“We ended up coming up with new ways of speaking to avoid that kind of moderation, and then we ended up embracing some of those words and they became common vernacular,” said Diaz of the University of California, Los Angeles School of Law. “It was all born out of this effort to resist moderation.”
This does not mean that all efforts to stamp out bad behaviour, harassment, abuse, and disinformation are futile. But Jarir says it is the root issues that need to be prioritized. “Aggressive moderation will never be a real solution to the damage we see from the business practices of big tech companies,” she said. “This is important for policy makers and to create better things, better tools, better protocols, and better platforms.”
Ultimately, she added, “You’ll never be able to cleanse the internet.”