Self-Censorship and Internet Double-Speak 

“My friend became sewerslidal after she was 🍇 by a PDF, and so she un-alived herself with a pew-pew.”   

If the sentence above seemed nonsensical, disrespectful, or dystopian, you are not alone. With increasing censorship by social media companies, many online creators who discuss true crime have taken to substituting serious words which may be flagged by algorithms. These terms, called algospeak, are euphemistic in nature, and immature in expression. Here is a list of common substitutions:   

Gun = pew pew   

Kill = unalive  

Suicide = self-delete or sewerslide   

 Rape = grape emoji  

 Pedophile = PDF file.   

 Nazi = Yahtzee   

Prima facie, this self-censorship is disgusting. It is childish, undignified, and indifferent to the seriousness of the discussion. It is true that automated content removal is correlated to the language used in a given video, but there is a lack of evidence to suggest that this self-censorship is effective in evading algorithmic moderation. And with the advancement of AI and large language models, algorithms are capable of understanding the context of such words and identifying the topic of discussion. The prevalence of algospeak suppresses the discussion of sensitive topics and disrespects those involved. Internet linguist Adam Aleksic’s 2025 book Algospeak is an etymological analysis of these terms, their origin, application, and future. Aleksic’s interpretation of algospeak is not a rupture from linguistic trends, but rather another evolution of historical trends which are uniquely shaped by social media platforms. Social media's main objective is to monetize its content and maintain order so far as it retains advertisers and users. Reducing hate speech and illegal content and promoting safety are only additional benefits to maintaining a company’s bottom line. Twitter, now called X, is a perfect example of the financial consequences to removing the veneer of concern. After Elon Musk’s acquisition of twitter, he changed the app’s name and eliminated nearly all moderation of its users’ content. Twitter had already been home to lots of explicit content including porn, violence, and death, but X now lacks almost any standard for removal. The app is home to unfettered white supremacists, Nazis, hate speech, and is some cases, even CSAM. The introduction of X’s own in-app AI Image generation was a tool for creating this content. Some of X’s biggest advertisers ended their business with Musk, and now the app’s Ad space and company evaluation, have plummeted in value. Immediately Following the murder of Charlie Kirk, videos of his public death circulated on X. Videos like this would previously have been relegated to less popular sites that are designated for unrestricted content, such as LiveLeak. Instead, anyone with the X app installed can view unblurred videos, from multiple angles, of a man’s neck pouring blood after being shot. Regardless of people’s sympathy, or lack thereof, for the political pundit, the virality of gore should be cause for concern. We must reevaluate social medias impact on violence and our perception of it.    

While terms like un-alive have a literal connection to the word be substituted—death and killing do result in the end of a person’s status as alive—other terms like pew-pew, grape emoji, or PDF File, are connected by onomatopoetic or phonetic similarity. These negatives facets of algospeak are unlikely to go away, and it’s an issue with both online platforms and those who use it. It trivializes the topics of discussion, it obscures reality and reduces empathic response. It’s both an indictment on platforms which censor important conversations and the way in which users have gone about finding alternatives. I think that ultimately, social media platforms fail as a medium through which these discussions can take place because we are expected to keep our conversation advertiser friendly. To the uninitiated, the use of these terms is insulting. Imagine a friend confiding in you about their “sewerslidality” or “grape emoji” experience. Imagine a doctor who tells a patient's family their loved one unalived. Abbreviations prove most effective in preserving connotation and meeting the goals of algospeak. SA and CSAM are identical stand-ins for their full counterparts, that is, Sexual Assault and Child Sexual Abuse Material. It is futile to condemn specific users that employ algospeak or demand that platform cease automated content removal. At the very least, it’s worth acknowledging both algospeak’s absurdity and practical applications.