Technique Spotlight: Algospeak

One way that online platforms can tell if people are spreading harmful content is by scanning the text for particular keywords. Basically, if you post something like “I’m going to att*ck and k*ll this person” on Social Media, a moderator would probably be alerted pretty quickly.

You might notice I wrote “k*ll”, not “kill”, because I don’t want scary moderators to come for my website. I used algospeak to try to make it harder for automated detection of sensitive keywords to succeed. Some marginalised groups use this technique to avoid online harassment from hate groups, some use it to enable the undetected spread of false information. Algospeak isn’t always used maliciously, but sometimes it is!


Real World Examples

Emoji Help Anti-Vaccine Posts Avoid Moderation on Facebook

Anti-vaxxers, whose activity is limited on many platforms, have started using elaborate quotes and phrases to spread cloaked misinformation about the drugs that help fight Covid-19. In Facebook groups and pages with names like “Died Suddenly,” users congregate and claim to be mourning friends and family members who have “eaten the cake,” which is code for taking vaccines. 

By Daniel Zuidijk on 19 Oct 2022
Tracking Exposed: Demanding That the Gods Explain Themselves

In the early days of online chat, AOL users used creative spellings to get around profanity filters, creating an arms race with a lot of collateral damage. For example, Vietnamese AOL users were unable to talk about friends named “Phuc” in the company’s chat-rooms.

By Cory Doctorow on 02 May 2022
Internet ‘algospeak’ is changing our language in real time, from ‘nip nops’ to ‘le dollar bean’

Algospeak refers to code words or turns of phrase users have adopted in an effort to create a brand-safe lexicon that will avoid getting their posts removed or down-ranked by content moderation systems. For instance, in many online videos, it’s common to say “unalive” rather than “dead,” “SA” instead of “sexual assault,” or “spicy eggplant” instead of “vibrator.”

By Taylor Lorenz on 08 Apr 2022

Placing Algospeak in an Influence Operation

Algospeak would likely be deplyed when Producing inauthentic content, or in response to moderation efforts during Calibration.

DISARM’s phases with DDB’s stages

Relevant Content from the Article Archives