Chekkee stands in unison to support our friends, colleagues, brothers, and sisters in Ukraine. #WeStandWithUkraine
The availability of different social networking platforms has broken down all barriers concerning communication. It paved the way for the people to connect and communicate seamlessly and almost instantaneously online with one push or click of a button. However, as these social networking communities start to flourish with subscribers, the chances of members getting too “comfortable” with the system become higher, thus the level of colloquialism or the informality of language also becomes extensive. Swear words, euphemism, hate speeches, discrimination, sexual-related expressions, and other foul terms are being utilized inside online messaging platforms and discussion boards. Though these expressions can be normal for some, the majority of internet users tend to get offended and it eventually breeds conflicts between users. This is where a profanity filter comes into action. A profanity or language filter works as an auto-correct (also called automod or automoderators) software subsystem which alters, modifies or removes terms that are deemed too offensive to read by both community admins and members. It can cover a wide variety of communication channels such as public chat rooms, community chat rooms, and online gaming chat rooms. However, some people are persistent with their unbecoming habits. Swearing becomes the new normal while inside a particular chat room and violators find ways to bypass or evade active profanity filters. They do this by using “tricks” to modify the words they’re using so that the profanity filter in place permits them one way or another. These deceptive and covert practices are proof that using a profanity filter alone is not enough in order to stay on top of your site’s moderation and protect your business reputation. It is essential to know and uncover the tricks that help people bypass profanity filters. That said, here are six common tricks used to avoid profanity filters.
1. Word Separators
First on our list is called word separators. This is a complex type of trick that can be used to ignore profanity filters. It involves the use of — separators (hence its name) such as spaces, periods, commas, and other non-alphabetic characters between each letter or character of a whole word without compromising its readability. Check out these examples:
- S…..M…..I…..L…..E
- S*M*I*L*E
- S m i l e
- S…m*i….L,e
These are good examples of words processed with separators. As you can see, there may be some unnecessary elements incorporated within each letter but it did not lose its readability. But, while non-alphabetical characters are quite effective in getting around profanity filters, letters can also be used as separators in some instances, making it even difficult for the profanity filter to recognize these violations. Take a look at these examples:
- SaaaaMaaaaIaaaaLaaaaE
- She Mosies In Late Everytime
Next, check these examples:
- His ex.
- Go bash it!
At first glance, it will take you a while to see the profanity in each phrase. But if you look more closely at the last 4 or 3 letters and read the characters on their own, you’ll be able to see the profanity present in the phrase. This is due to the fact that profanity filters are sometimes not intelligent enough to distinguish which word should be filtered and which should not. In the end, it identifies some sentences as false positive statements. The best course of action to prevent your profanity filter from missing such tactful letter arrangement is to upgrade your software to its latest patch and be acquainted with these types of non-harmful texts. In response, you come up with plans on how you will integrate your filter standards to address these kinds of texts.
2. Embedding Words In Other Words
Have you ever heard of the “Scunthorpe Problem”? It is based on a small town in England called Scunthorpe, which unintentionally ran afoul of Facebook’s automated moderating filter. The latter deemed the English town’s name as a term or text depicting profane words. Due to this, locals from this town had a hard time advertising their services on Facebook. Since then, the Scunthorpe Problem became a blanket term for profanity embedded in all proper names and dictionary words. The following are good examples of names categorized under Scunthorpe Problems:
- Butts
- Cummings
- Medick
- Dickinson
- Suconcock
It left a huge impact especially for brands where these terms are being incorporated. A cascade of unintentional blocking of websites, emails, forum posts, or search results by an automated profanity filter ensued simply because these texts, albeit unintentionally offensive, contain letters that when arranged in a certain manner are classified as profane, obscene, or outright vulgar. Users, on the other hand, take this as an advantage to cloak their profane texts by embedding them into other words. Here’s a concrete sample:
- qwertySMILEuio123
Or, they can go for a more complex tactic by combining phonetic replacements, separators, or leet speaks:
- abcd$M1LEabcd
- abcdSMYLEabcd
- abcdSSSMMMIIIILLLLEEEEabcd
- abcdSM||_Eabcd
- abcdS*M*I*L*Eabcd
With the presence of this technique, it pays to have your automated profanity filter be smart enough to not only distinguish custom-embedded words but also possess the ability to identify when these embedded words are not meant to be harmful. Smarter scrutiny of these terms prevents the same fate of the town of Scunthorpe from recurring.
3. Phonetic Letter Replacements
This is the process of simply replacing characters with other alphabetic characters. In some instances, characters within a word can be strategically removed without losing the phonetic structure of the word. Here are some examples:
- SMYLE/SMYAL/SMYLLE
- THIC/THICC/THICCK
- ATTAK/ATACK/ATTACC
As you can see in example number 1, the letter “I” is replaced with “Y” as this can be a practical alternative without losing its phonetic structure. As for examples 2 and 3, implementing the removal and addition of characters altogether is used to render the terms to look different without losing its phonetic structure. Phonetic letter replacement is one of the most commonly used tactics to get around profanity filters, as most profanity filters are not tailored to recognize and understand phonetics.
4. Repeat Characters
The name itself explains how repeat characters tactic works. It mostly involves the over repetition of characters in a word to hide their message. Check out these examples:
- sssssmmmmmmiiiiiiilllllllleeeeee
- hhhhhhhhheeeeeeeelllllllooooooooo!
Although it looks quite simple, this tactic more often than not bypasses profanity filters. To protect your online community or chat room from being flooded by these types of words, it is essential to set up your profanity filter to detect multiple occurrences of the same character. Failure to do this will allow users to seamlessly sidestep profanity filters.
5. Grawlix
By definition, a Grawlix is a string of typographical symbols used as a substitute to cover profane, offensive, or obscene words. This is usually be seen in comics strips. It is a group of non-alphabet characters put together that clearly identifies a profane word. Here are some examples of Grawlix:
- &@$#%! What have you done?!
- You’re such a @&*^!@!(&(!
- Whatever, ^@%!$
At first glance, Grawlix might not look offensive compared to the other tactics cited above, but this is a very effective strategy for members to express profane emotions without triggering your profanity filter.
6. Leet Speak
Last and definitely not the least in our list is called Leet or “L33T” speak. It is another trick that involves replacing certain letters or characters with other characters that usually resemble or are closely similar to the original character. This is a very effective way to hide from content filtering as most traditional profanity filters are not well-integrated to recognize non-alphabetic characters. Here are some examples of Leet Speak:
- $@|_|_¥ in place of Sally
- |)0N’T $M1|_E in place of DON’T SMILE
To make this easier to comprehend, here is a chart of commonly used characters in the Leet Alphabet:
Leet is a very complex and tricky way to evade any profanity filter as most profanity filters don’t work well with non-alphabet characters. Essentially, it is best to integrate your profanity filter not only to keep pace with the most updated ways to bypass filters but also to easily identify whether a symbol is used correctly or not. Doing so avoids the likelihood of UGC-related mishaps that can affect your brand in the long run.
Raise Awareness, Upgrade Your Profanity Filter Here And Now
The internet sphere is getting smarter. As time passes, newer ways to attack filters will be developed by users that will allow them to say whatever they want, whenever, and wherever they want to. It is best to set up an efficient and dependable software that can effectively moderate and filter any harmful, vulgar, or misguiding user-generated content. Chekkee offers top-notch moderation services that can cater to online platforms of all sizes. With moderation services with a touch of both human and artificial intelligence, our expert moderators ensure that your brand is 100% protected against unnecessary and inappropriate user-generated content. Want to know more about how we take human content moderation to the next level? Contact us today, and we will be more than happy to make assessments!
