Algospeak summary

Euphemism is one way in which content creators can avoid the axe of moderation. But there are many others.  

People have been searching for creative ways of avoiding censorship from time immemorial. Norman Mailer replaced the f-word with “f-u-g” in The Naked and the Dead to navigate publishing restrictions. Going from the profane to the sacred, we have such ancient examples as the stylized fish symbol, often seen now on the bumper stickers that Christians use to signal their faith. This originated in the period of early Christianity when believers used it to identify each other in an era of Roman persecution. 

What’s the first internet-specific term you ever heard? If you’re in your thirties, it might have been “pwned” from gaming forums in the early 2000s, or perhaps “rickrolled” from that infamous YouTube prank involving Rick Astley’s “Never Gonna Give You Up.” If you’re a bit older, you might remember first encountering “LOL” in early chat rooms or hearing someone described as a “newbie” on bulletin board systems. These terms didn’t emerge from corporate marketing departments or viral campaigns; they grew organically from the communities that used them.

Before roughly 2004, the internet was fundamentally decentralized, scattered across message boards, forums, and early websites. The next wave of the internet introduced massive centralization through platforms like Facebook, Twitter, and YouTube. These social networks connecting millions of users within unified ecosystems allowed new terms to spread faster than ever before. Yet even these early social platforms operated on relatively simple, transparent systems that users could understand and navigate.

Most people know memes as humorous images with bold text – from the early “I Can Has Cheezburger?” cats to the so-called Wojaks, those simple black-outlined drawings of bald men expressing various emotions. But did you know the term “meme” has its roots in evolutionary biology?The term was coined by biologist Richard Dawkins in his 1976 book The Selfish Gene. Dawkins derived it from the Greek word mimema, meaning “something imitated,” and deliberately shortened it to echo “gene.” In doing so, he drew a fascinating parallel between how genetic information replicates through biological evolution and how cultural information spreads through human societies.

In Dawkins’ original conception, memes encompass any units of cultural transmission – ideas, behaviors, or vocabulary – that jump from person to person through imitation. This biological metaphor reveals something about how language evolves: words and concepts function like viral organisms, as if us humans are mere hosts among which they spread.

TikTok’s algorithm represents a quantum leap in content recommendation. Users often report the uncanny experience of the app practically seeming to read their minds – anticipating their interests before users are even aware of them. The platform’s success depends on its ability to track user behavior with remarkable precision – monitoring not just what users watch, but how long they do so, what they skip and when, and many other behavioral signals. Using advanced machine learning (or else some kind of black magic), TikTok’s algorithm optimizes for maximum user engagement, keeping people scrolling for hours.This algorithmic sophistication creates powerful incentives for linguistic trends. When a new slang term emerges, users who engage with it trigger the algorithm to suggest more content featuring the same term. This only incentivizes creators to jump on trendy terms. TikTok’s audio-reuse feature amplifies this effect, allowing sounds and phrases to spread across thousands of videos with minimal friction.

The word “rizz” – short for charisma – exemplifies this phenomenon. Emerging in 2023, “rizz” spread rapidly through TikTok’s recommendation engine. The viral explosion reached its peak in October 2023 with the “Rizzler song,” which cannily crammed multiple trending Gen Alpha slang terms like “rizz,” “sigma,” and “skibidi” together into deliberately nonsensical lyrics.

Goofy new slang terms are all well and good, but there’s more at stake. While algorithms were designed to connect and entertain people, they have some serious unintended consequences. Let’s look at them.

In 2010, Wharton professors conducted an analysis of New York Times articles – specifically, how and when users shared them. They discovered a troubling pattern: the more emotionally charged a piece of content was, the more likely it would be shared. Notably, it didn’t matter whether the emotion was positive or negative; what mattered was its intensity. This finding has been replicated again and again, across multiple formats and platforms.