Detect and delete Spanish posts

Status
Not open for further replies.

Deleted member 737209

Well-Known Member
Newcomer
Joined
Jan 28, 2025
Messages
83
Reaction score
56
Trophies
0
XP
215
There are many new users making Spanish posts which quickly get deleted. They seem to be scams directing consumers to fake customer support numbers which are actually scam call centers. Not sure exactly how to implement it but it would be something like this.

Send post to Google Translate
If detected language is Spanish
Then
-Block post
-Suspend account
-Flag account for manual review
Else
-Do nothing
 
"Vaya! Pero cómo no lo había pensado antes!?"
- A mod... who learned Spanish just by deleting spam written in mexican.

Thanks for the tutorial though, sadly I'm not a mod.
5/5 would attempt when I'm a mod.
 
Spammers will find a way to evade guards in place to prevent things like this from happening. It's not rocket science and neither robust in every way.

Second, is that GBAtemp has a rule that does allow languages other than English in the dedicated language groups, but with an exception for other forums.
 
Spammers will find a way to evade guards in place to prevent things like this from happening. It's not rocket science and neither robust in every way.

Second, is that GBAtemp has a rule that does allow languages other than English in the dedicated language groups, but with an exception for other forums.
In those spam posts, there's always a phone number provided. If submissions with numbers in a specific format get flagged for moderator review, it may prevent some of these posts. At that point, the spammers would have to detect and change how they format the number, leading to a perpetual cat-and-mouse game.

That would be my suggestion, though I understand it may hold up legitimate posts containing similar numbering formats, such as Pro Action Replay codes, hex, etc.

To be honest, the mods are really fast at deleting the spam. There has been a recent increase in spam posts, but it seems to come and go in waves.
 
Please no automated data processing by Google (or any other third party) getting all newbie posts. Not that Google would be very fond of getting an abundance of posts for translation putting load on their servers.
Even a local plugin could probably handle language detection – the question is why. As long as a English translation is provided other languages are often accepted and there are foreign language groups as well.

The spam filter probably works with a word list and regular expressions. No wonder Spanish posts sneak by the filter at the moment as many common phrases in English have been blacklisted for newbies. Sadly the more strict the admins adjust the spam filter the more false positives will occur. There are false positives already: Newbies sometimes can't quote posts containing certain keywords or even links.

I wouldn't hold my breath for the idea that any filter or even an LLM could permanently solve the spam problem. The will try again and again. Replacing letters and number with similar characters from Unicode, obfuscate phone numbers and other characteristics. Only human (pre-)moderation would get rid of all spam but that would require an insane amount of work and cause unpleasant delays for newbies. For the most the filters worked in the last few months… after all the "Bitcoin Recovery", "Black Magic", "Hire a hacker", "Airline support number" and whatnot in English.

We normal members can't do more than report spam.
 
Status
Not open for further replies.

Site & Scene News

Popular threads in this forum