Hello, dear reader. I am going to talk about something you are almost certainly quite familiar with. The companies that control social media sites have long implemented policies restricting the content that can be viewed as offensive or distressing. To get around this obstacle, users of these platforms have coined terms with the same meaning as words forbidden by the company’s policies. A rather (in)famous example of this is the term unalive replacing the word kill. This practice, now widely known as AlgoSpeak, sometimes goes outside of the online landscape and gets used in everyday language.
While this may seem like a harmless evolution of language, it is actually extremely dangerous because AlgoSpeak terms are euphemisms that lead to things being taken less seriously. They also lead to restrictions on expression being normalized, leading to the suppression of information and people being unable to share ideas that challenge the dominant narrative. This will ultimately harm society, allowing established institutions to be free from scrutiny and preventing people from being exposed to different views than those to which they are used.
How AlgoSpeak Was Born
The phenomenon known as AlgoSpeak began when users of social media attempted to get around the terms of service of different social media companies that censor certain words and suppress certain topics (Kreuz, R. J., 2023, April 13). While social media allows common people to have a platform and express their beliefs and share their thoughts with millions of others like them to an unprecedented degree, these platforms are controlled by companies focused on maximizing profit and minimizing risk, leading to certain words and topics being banned or heavily restricted. To get around these regulations, people have created AlgoSpeak words and phrases that communicate the same messages as the censored words (NPR Podcast, 2025, July 20).
I am almost certain that you have seen instances of this in videos or comment sections. The reason why these AlgoSpeak terms can be used despite meaning the same thing as the banned words is that the censors on social media sites are not people, but programmed algorithms that cannot recognize meaning, only words. This means that these algorithms will be triggered if they recognize a word that is forbidden, regardless of the context in which it is used, and will fail to recognize when a replacement word is used to communicate the same message as the banned one.
When Euphemisms Become Mandatory
These policies have a very damaging effect on modern discourse by trivializing issues and normalizing censorship. The first way modern discourse has been negatively impacted is through the promotion of euphemistic language. AlgoSpeak makes the topics being discussed seem less serious than they are, leading to things that would otherwise cause outrage to be seen as mundane (Plotiar. 2026, April 5).
This is harmful to society and social discourse because it gives people a false impression of reality and prevents people from accurately analyzing the world. It also pacifies the public by making things that would otherwise stir people to action seem innocuous. (Repkin, J. 2023).
You may point out that euphemisms have always existed and that this is nothing new. While you’d be correct, the rise of censorship by social media companies has essentially made AlgoSpeak mandatory to have access to a large platform. This normalizes euphemistic language in an unprecedented manner.
Who Really Controls What You See
You still may not be entirely convinced of the danger; after all, don’t companies have the right to decide what kind of content is allowed on their platforms? While this may seem like a reasonable attitude, it is actually incredibly dangerous. The first danger of this policy is that it allows large companies to determine what kind of ideas and information are allowed. The power that big business has over the spread of information and ideas is even greater than that of governments today (Heins, M. 2014).
The censorship these businesses are implementing is done in the name of advertiser-friendliness the same environment that made AlgoSpeak necessary in the first place. This means that these companies will have an interest in suppressing truthful information that would be of benefit for the public to know if it paints advertisers in a bad light.
This is because social media companies are primarily interested in earning a profit, and therefore will seek to suppress information if such information being available damages their profits. This means that truth and the spread of accurate information take a back seat to avoiding controversy and catering to advertisers’ needs.
When Governments and Corporations Align
Another big danger of this corporate censorship is that it allows governments to crack down on dissent. Something which I am almost certain you do not want for this country. Examples of this include the US government collaborating with social media companies to suppress voices, or countries demanding that social media platforms forbid people from discussing topics critical of their country or government (Petti, M. 2025, October 13).
It is also very dangerous because it allows institutions that are not accountable to the public, as shareholders rather than citizens elect their heads. These two factors mean that while social media has the potential to empower ordinary people by allowing them to share ideas and beliefs on a larger scale than was possible in the past, the infrastructure is controlled by governments and conglomerates, which are very often controlled by elites rather than regular people. AlgoSpeak is ultimately a symptom of this imbalance of power.
Why Even “Reasonable” Censorship Is a Problem
I imagine that while you may believe in freedom of speech as a general principle, you likely believe there are certain cases where certain opinions need to be censored because they spread misinformation, dehumanization, or extremism. Even in seemingly reasonable examples of censorship, such as suppressing hate speech, they are dangerous because they allow hateful ideas to go unchallenged through lack of exposure and allow hate groups the ability to claim they are being persecuted, proving that their theories are correct (Ben-Ami, D. 2022, March 31).
Even in cases where censoring hateful or bigoted speech does reduce its spread, it is still dangerous to censor it because our perceptions of speech being harmful and untrue (the perceptions that make censorship otherwise acceptable to those who favor free speech) are often influenced more by emotional reactions than clear-headed reflection (Kubin, E., Christian von Sikorski, & Gray, K., 2024).
Also, censorship often tends to push those being censored into echo chambers and radicalize them, especially top-down, centralized censorship (E. Lane, J., McCaffree, K., & Shults, F. L., n.d.).
What We Can Actually Do About It
This may strike you as something that cannot be solved and therefore is something not worth discussing. This would be quite mistaken, both because it would be to abandon thinking and the idea that nothing can be done is false.
The first thing that can be done is to build and embrace uncensored social media. While these platforms are generally associated with the far-right, the concept is still important for maintaining open discourse in the online world and reducing our dependence on AlgoSpeak as a workaround.
The next thing that can be done is to embrace user control of algorithms (Malik, A. 2026, June 17) or get rid of algorithms entirely to avoid companies and other institutions controlling what people are exposed to (Rawly, 2026, May 26).
Finally, replacing robots with actual humans to enforce the terms of service might also help, as humans can understand and interpret meaning. This means that they can understand the context of what is being talked about and therefore will be able to more accurately assess whether or not content is genuinely harassing or threatening or is simply speaking on a topic which may cause discomfort or controversy.
All of these solutions, while imperfect and not sufficient on their own, will help to move social media into a more open direction and give people a greater ability to spread information and ideas without having to resort to AlgoSpeak.
Works Cited
Ben-Ami, D. (2022, March 31). The Weimar fallacy. Radicalism of Fools.
Heins, M. (2014). THE BRAVE NEW WORLD OF SOCIAL MEDIA CENSORSHIP. Harvard Law Review Forum.
Petti, M. (2025, October 13). Algorithmic censorship changes the way we talk. Reason.Com.
Plotiar. (2026, April 5). Euphemism. Plotiar.
Rawly. (2026, May 26). Social Media Apps Without an Algorithm in 2026.
Repkin, J. (2023). Euphemisms are powerful — and dangerous. The Mesa Press.
Free guide when you subscribe
Enjoyed this? Get Say It Out Loud, free.
Our menopause conversation guide: what to say to your doctor, your therapist and your partner, plus a symptom tracker. Subscribe free and it lands in your inbox, with $10 off the shop and new stories twice a month.
No spam. Unsubscribe in one click.
