It is pretty safe to say that censorship is a controversial issue, as it can limit what can be said and expressed in public and private spaces, including the internet. Most baseline forms of censorship can block universally illegal content such as extreme violence, but in many areas globally, this may be employed to suppress opposition voices, or prevent certain issues from gaining public traction. While we are pretty familiar with censorship that is perpetrated by governmental authorities, there is also another type of censorship that has taken an interesting form online.

If you have been perusing social media or content sites such as TikTok, Twitter, and YouTube in recent years, there has been an increase in bleep outs or certain alternative vocabulary used when discussing either morbid or controversial terms. Even themes as seemingly benign as ‘link in bio’, and dealing with obesity, may be rendered as alternate forms as well. This phenomenon that we have been seeing in the 2020s is pretty reminiscent to the one I have seen in Chinese online spaces since the early days of the internet. I have done a little introduction to the Chinese side of this, through the use of subversive puns such as ‘Grass Mud Horse’ and ‘harmonisation’, so check that out for more context.

As such, seeing some of these alternative vocabulary being used in the English-speaking internet has piqued my interest, not just because it touches upon the issue of censorship, but also to see how internet users in the anglosphere try to sidestep certain topics in their social media and content uploads.

On social media like Twitter, Facebook, TikTok, and YouTube, there is this certain mythical being called the ‘algorithm’, which ‘decides’ the content to push towards potential viewers based on viewing habits, interests, and whatever personal, browsing, or viewing data they can scrape off the user. This ‘algorithm’ can also suppress certain controversial topics, such as those discussing war, violence, or certain adult activities that may be disturbing or distressing to some users. Yet, for people whose livelihood depends on the success of their content reaching out to users, which can be converted to views, monetisation, and interactions, the drive to maintain this success and be seen pretty much involves them to ‘appease the algorithm’, and prevent getting suppressed by the ‘algorithm’.

Censorship on these social media sites takes the form of content moderation. Topics such as risky activities that can pose a danger to life, sexual themes (since the minimum age to use these sites is 13 across most countries, and 16 in Australia), and gruesome imagery would be generally deemed to go against moderation algorithms, and these types of content, as with users who share these types of content, would generally be penalised by limiting their visibility to other users, or on YouTube and TikTok, where content monetisation is a thing, limiting the ability of these types of content to be seen and monetised.

For creators who share educational topics like history, crime, politics, morbid topics, and the ‘dark side of the internet’, there would be a need to sidestep this form of censorship to be seen. And so, enter this thing called ‘algospeak’.

Algospeak is a type of avoidance speech that netizens use to evade content moderation filters, and I am pretty sure like in the Chinese internet and the anglosphere internet, similar forms of algospeak occur in other languages as well. But anyway, it can manifest as the bleeping out of certain sensitive words which may trip filters, or the replacement of such terms with those that can carry the meaning across, but may occur as euphemisms, similar-sounding words, or the use of emojis. Perhaps one of the most famous algospeak terms is the use of the word unalive to describe things like ‘killing’, ‘death’, or if used with a reflexive (i.e. unalive oneself), ‘suicide’, and the word grape or the grape emoji to describe sexual assault or rape.

The use of the terms grape, corn, and acoustic follow the pattern of using some form of rhyming slang to obfuscate these sensitive terms. This pattern has been seen in Cockney rhyming slang (e.g. using Tom, from tomfoolery, to mean ‘jewellery’), though the origins and supposed original intentions are generally different from algospeak. Some terms also have the added layer of deriving from leetspeak, another type of internet slang where other characters may replace letters or letter sequences in the English alphabet (like $ to replace ‘s’). One example is the term le dollar bean, deriving from the leetspeak le$bian to mean ‘lesbian’.

Others seem a bit more clandestine. For example, some short-form videos have used the term accountant to mean ‘sex worker’, and mascara to mean ‘sexual assault’. The origins of these algospeak terms are very unclear to me, though I would like to learn more about their etymologies. Perhaps this is part of a continual arms race between social media users and the evolving content moderation methods and systems.

While the origins of algospeak is not really clear, its use has exploded with the rise of TikTok and the proliferation of its content moderation systems. Some pressures to maintain or the expand the use of algospeak is the inconsistent or biased policing of certain content moderation, or the lack of consideration of context when moderating content. Despite this (and YouTube) being the most well-known social media space where algospeak is used, algospeak has also been used in some other controversial groups as well, such as those promoting harmful behaviours or spreading misinformation. On the flipside, support groups that provide aid and guidance on dealing with substance abuse, traumas, and self-harm may also be forced to use algospeak to prevent their activities from being censored, and may view the use of algospeak as disrespectful to the people who have experienced these traumas.

As content moderation expands in use not only on social media, but also in other forms of media, it seems that algospeak can only evolve in use, bringing in more innovations over time. In addition to Gen Alpha slang, maybe we would see more algospeak terms used in the media we consume in time to come.

Further Reading

Algospeak, How Social Media is Transforming the Future of Language by Adam Aleksic, published by Penguin Random House in 2025