‘Algospeak’ changes our language in real time

“Algospeak” is becoming more and more common on the Internet as people try to bypass content moderation filters on social media platforms like TikTok, YouTube, Instagram, and Twitch.

Algospeak refers to code words or phrases that users adopt to create a brand-safe dictionary that will prevent their posts from being removed or de-ranked by content moderation systems. For example, in many online videos it is common to say “not alive” instead of “dead”, “SA” instead of “sexual assault” or “spicy eggplant” instead of “vibrator”.

As the pandemic forces more people to communicate and express themselves online, algorithmic content moderation systems have had an unprecedented impact on the words we choose, particularly on TikTok, resulting in the emergence of a new internet-focused form of Aesop.

Unlike other mainstream social platforms, the primary way content is distributed on TikTok is through an algorithmically selected “For You” page; having followers is a secondary metric that doesn’t guarantee people will see your content. This change has caused average users to adapt their content primarily algorithmically rather than following it; This means that it is more important than ever to comply with content moderation guidelines.

When the pandemic broke out, people on TikTok and other apps started referring to it as “”.Backstreet Boys reunion tour” or as platforms, “panini” or “panda express”, listing videos that mention the pandemic by name in an attempt to combat misinformation. When teens began discussing tackling mental health, they talked about “being dead” in order to be able to have candid conversations about suicide without algorithmic punishment. Long censored by moderation systems, sex workers call themselves “accountants” on TikTok and use the corn emoji instead of the word “porn”.

More users are twisting their tongues as discussions about important events are filtered by algorithmic content delivery systems. Recently, when discussing the invasion of Ukraine, people on YouTube and TikTok used the sunflower emoji to denote the country. Users will say “blink in lio” for “link bio” while encouraging fans to follow them elsewhere.

Aphorisms are especially common in radicalized or harmful communities. Pro-anorexia eating disorder communities have long embraced variations on moderate words to avoid restrictions. A paper from the Georgia Institute of Technology’s School of Interactive Computing found that the complexity of such variables increases 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 code words, calling vaccinated people “swimmers.”

Adapting the language to avoid scrutiny takes precedence over the Internet. While many religions avoid mentioning the devil so as not to invoke the devil, people living in oppressive regimes have developed code words to discuss taboo topics.

Early Internet users used alternative spelling or “leetspeak” to bypass word filters in chat rooms, picture boards, online games, and forums. But algorithmic content moderation systems are more common on the modern Internet and often silence marginalized communities and important debates.

During YouTube’s “adpocalypse” in 2017, LGBTQ creators talked about de-monetizing videos for saying the word “gay”, as advertisers withdrew their money from the platform for fear of unsafe content. Some have started using the word less or replacing others to continue monetizing their content. More recently, TikTok users have started to say “abundance” rather than “homophobia” to indicate they are LGBTQ, or say they are members of the “leg booty” community.

“There’s one line we have to watch out for, it’s a never-ending battle to say something and try to get the message across without saying it directly,” said Sean Szolek-VanValkenburgh, a TikTok creator with over 1.2 million followers. “This is disproportionately affecting the LGBTQIA community and the BIPOC community because we are the ones who make the talk and prepare the colloquia.”

Kathryn Cross, 23-year-old content creator and founder of Anja Health, a startup offering umbilical cord blood banking, also said that conversations about women’s health, pregnancy and menstrual cycles are consistently low on TikTok. She replaces the words “sex”, “period” and “vagina” with other words or spells it with symbols in the subtitles. Many users say “nipples” instead of “nipples”.

“It makes me feel like I need a disclaimer because I think using these oddly spelled words in your captions makes you look unprofessional,” he said, “especially for content that needs to be serious and medically inclined.”

Because online algorithms often mark content that mentions certain words as context-free, some users avoid using them altogether because they have alternative meanings. “You literally have to say ‘salty’ when talking about crackers right now,” said Lodane Erisian, Twitch creators’ community manager (Twitch takes the word “cracker” as an insult). Twitch and other platforms have gone so far as to remove certain phrases because people were using them to convey certain words.

Black and trans users and those from other marginalized communities often use algospeak, replacing the words “white” or “racist” to discuss the oppression they face. Some are too nervous to utter the word “white” and simply hold their palms towards the camera to indicate White people.

“The truth is that tech companies have been using automated tools to moderate content for a very long time, and even though this is touted as complex machine learning, it’s often just a list of words they think is problematic,” said Ángel Díaz. UCLA School of Law, which studies technology and racial discrimination.

In January, linguists Kendra Calhoun and Alexia Fawcett gave a TikTok presentation for an event on language hosted by the Linguistic Society of America. They summarized how new algospeak codewords came to be by auto-censoring words in TikTok’s titles.

TikTok users now use the phrase “le dollar bean” instead of “lesbian” because that’s how TikTok’s text-to-speech feature pronounces “Le$bian”, a censored way of writing “lesbian” that users believe will avoid content moderation.

Evan Greer, director of Fight for the Future, a nonprofit digital rights advocacy group, said that trying to eliminate certain words on platforms is stupid business.

“One, it doesn’t actually work,” he said. “People who use platforms to organize real damage are pretty good at figuring out how to circumvent these systems. Second, it causes collateral damage to real speech.” Trying to organize human speech on the scale of billions of people in dozens of different languages, and struggling with things like humor, sarcasm, local context, and slang, Greer argues, cannot be done by simply ranking certain words down.

“I think this is a good example of why aggressive moderation will never be a real solution to the damage we see from the business practices of big tech companies,” he said. “You can see how slippery this slope is. Over the years, we’ve seen more and more misguided demand from the general public for platforms to quickly remove more content, no matter the cost.”

Major TikTok creators have created Google documents that are shared with lists of hundreds of words that they think are problematic for the app’s moderation systems. Other users keep a tally of terms they think are restricting certain videos by trying to reverse engineer the system.

“Zuck Got Me For,” a site created by a meme account manager run by Ana, is a place where creators can upload nonsense content that is banned by Instagram’s moderation algorithms. In a manifesto about his project, he wrote: “Creative freedom is one of the only silver linings of this flaming online hell we all live in… As algorithms tighten, independent creators suffer.”

It also outlines how to speak online to avoid filters. “If you have violated the terms of service, you can use profanity or ‘hate’, ‘kill’, ‘ugly’, ‘stupid’ etc. You may not be able to use negative words such as “I often write, ‘I opposite of love xyz instead of ‘I hate xyz’.

The Online Creators’ Association, a labor advocacy group, has also released a list of public requests asking TikTok for more transparency about how it manages content. “People have to turn down their language so as not to offend these all-seeing, all-knowing TikTok gods,” said Cecelia Gray, TikTok creator and co-founder of the organization.

TikTok offers an online resource center for creators who want to learn more about their recommendation system, and has also opened multiple transparency and accountability centers where guests can learn how the app’s algorithm works.

Vince Lynch, CEO of IV.AI, an AI platform for understanding language, said that in some countries where moderation is more severe, people are creating new dialects to communicate. “They are becoming real sub-languages,” he said.

But as algospeak becomes more popular and substitute words become common slang, users are discovering that they need to be more creative to avoid filters. “It’s turning into a game of hitting the mole,” says linguist and author of “Because the Internet,” Gretchen McCulloch, describing how the Internet has shaped language. As platforms begin to notice, for example, that people say “seggs” instead of “sex,” some users are reporting that they believe the substitute words are even flagged.

“We’re creating new ways of speaking to avoid this kind of moderation,” said Díaz of the UCLA School of Law, “and eventually we’re adopting some of these words and they’re becoming common vernacular. All of this arises out of an effort to resist moderation.”

This does not mean that all efforts to eliminate bad behavior, harassment, abuse and misinformation are futile. But Greer argues that these are the key 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,” he said. “This is a task for policymakers and to build better things, better tools, better protocols and better platforms.”

Finally, he added, “You can never sanitize the Internet.”