The problem with the censor filter
I understand the necessity of a censoring mechanism to prevent a sudden intake of a potentially triggering message, however this filter seems to be getting more and more invasive every day. How many times I've been trying to apologize from a member or a fellow listener who had sent me an offline message with a "Sorry I was asleep" reply and it strikes me with this:
"Something in your message appears to violate our terms of service (such as sharing personal contact information or social media accounts, or containing inappropriate content). 7 Cups depends upon you keeping your identity secure and private to protect everyone involved. Please also take care to show kindness and respect in every interaction here on 7 Cups."
At first I thought what info?? Then I realized that it considered the "asl" in the "asleep" as the archaic acronym of "A/S/L?" and was censoring my message. I guess the wordlist for the censor filter is a standard one, but who really uses asl anymore? What's more strange is sometimes it works, sometimes it doesn't work and it always accepts a.s.l instead of asl so it’s pretty useless.
The problem is even bigger in the use of the Turkish language. In Turkish we have;
Aslı - a common female Turkish name. One member I had who was trying to tell people her name kept getting bombarded with that "Your message includes contact info.." message she thought that 7 Cups somehow managed to probe into her personal details and figured out her real name and was trying to prevent her from sharing it with others. She thought she was doing something wrong until I told her the cause of the problem.
“aslında” - means "in fact," which as you can guess one of the most common words daily speech,
“asla” - means "never," which is also pretty common,
“aslan” - lion, leo the zodiac sign,
“aslına bakarsan,” “işin aslı,” “aslolan şöyledir” etc - different variations of the above since Turkish is an agglutinative language. So these are pretty common usages of figure of speech in daily language.
All of these get filtered by the censor and make talking very difficult in Turkish.
So how do we handle this as Turkish speakers? We write “as.lında” or “as.la” instead. We have to be very careful while speaking the daily language in Turkish as even a sentence like “In fact, I’d have never guessed that the truth was like that” triggers the censor filter 3 times in just 1 sentence.
An ill-intentioned user can just type “a.s.l?” and get away with it as well. However, a new user who doesn’t know the mechanics of the censor filter will be pretty surprised and blame themselves, if they can even figure out the existence of a censor filter in the first place. This makes the primary sufferers of the filter these new good-intentioned users.
Long story short, the censor filter is easily overcome. It only hurts those who are new and don’t know about it. It makes them question the anonymity and confidentiality of the chat (“Am I being watched here?”). As a listener of 15k chats I know how important it is for the member to be able to vent off freely. The natural flow of the communication shouldn’t get disturbed. When they get such a warning, it discourages them. Sometimes they write a whole paragraph and the censor filter deletes it. This is very dis.couraging for them. They even leave the chat because of that. They can’t see their original message either.
Problem: Censor filter is too invasive and disrupts the natural flow of communication.
My solution: Disable the censor filter.
Middle ground: At least disable the “asl” filter, as it seems that the number of normal communications it prevents is much more than the negative situations it’s trying to prevent, as almost nobody uses asl anymore.
Thank you for reading :)
Note: Even while trying to make this post I kept getting the “Something in your forum post sounded like you might be in crisis. If you are thinking about harming yourself or someone else, please refer to these crisis resources.” message, even though the post has nothing to do with self harm. I had to go over and over in my post to find out what it might have thought was self-harming, since there isn't even a list to check what the censor autofilters. I had to write over and over and that’s why I had to divide the post into several parts, to find out which part exactly got stuck in the filter. This only proves the amount of invasiveness of the filter.
@bluelotus99
Hii , I am sorry you're experiencing this , censors can be quite frustrating at times, I get it too, specially having to retype the entire long paragraph is certainly annoying and stressing , while it's increasing response time for listeners (having to retype from the start figuring out a way to rephrase the apparent censored word ) it definitely is discouraging for members too, it's not easy opening up and when experiencing these instances , where there's unnecessary prompts and entire messages being deleted , it can be even more off putting .
I am not sure this form is still in use, but could be a try ,I think ! It's a censor review form, if you'd like to send it out for a couple of mostly used words by you / or the members that could be wrongly censored 😊
HERE !
Tagging @7CupsCommunity if they could suggest another best practice for the same , thankyou ❤
@Sunisshiningandsoareyou
We couldn't say it any better than us.
@bluelotus99 7 Cups banned "asleep" and other words containing ASL once before. It was fixed in November 2018. It seems someone ignorant of this has recently made the same mistake again! The abbreviation for American Sign Language is also banned.
You're right to point out that the censor causes unexpected problems in other languages like Turkish. Dutch is another example. You can't say "hoe" (in English, "how") For example "Hoe gaat het?" (meaning "How are you?", literally "How goes it?") is censored.
Charlie
@RarelyCharlie
I personally think the censor is one of the best safety features on this site. I of course, sometimes accidentally trip it, and it's a bit inconvenient, but it does serve as a reasonable safety precaution. As it says in the site's safety factsheet, "Censor Reports are automatically triggered based on specific phrases". I think this feature could/should be expanded!
So while the censorship is 'annoying' during a chat, it's a necessary safeguard and hopefully helps keep vulnerable members safe. I really hope they expand upon it. The automatic reporting feature, and the details contained in that, could help streamline other site issues and concerns.
@lovelyArrow211
saying "ASL" to a vulnerable member isnt' going to harm them. the censor is failing.