New'ish type of spam I'm seeing

[Sal Collaziano]

Something I haven't seen until recently is new members signing up and after a few normal looking posts, start replying to discussions and adding link code in the quoted part of their responses. It's also setup so that it's not visible without viewing the code. So something to consider. Not sure how "new" this is, but I only recently had it become an issue...

 

[BassMan]

Not new and I hate those...

Also, a lot of human spammers with AI content. Hate them even more.

 

[Andy.N]

It's not new. New members, post right after signing up. Nothing suspicious but later they went back and edit their posts to insert url. Sometimes using color code to hide the url.

 

[Sal Collaziano]

It's not new. New members, post right after signing up. Nothing suspicious but later they went back and edit their posts to insert url. Sometimes using color code to hide the url.

Yes, that I've seen, but never to the actual quoted part of the message. Only new to me, I guess.

 

[BassMan]

Yes, be aware when someone replies to a very old thread (and quoting it).

 

[Sal Collaziano]

Yes, be aware when someone replies to a very old thread (and quoting it).

Yes, that's for sure...

 

[CedricV]

Very annoying. I've seen them coming on my forum too. No introduction topic, just straight reply to a subject. Nothing really suspicious at that time, but then after a few days they come back and edit a link in their original post.

 

[Rhody]

Yes we have had that for some time, too. Often they'll revive a thread from 20+ years ago with a seemingly AI post or comment. If you look at it carefully, it probably has a link in it. We don't allow editing after the first 2 hours.

 

[BassMan]

Yes, incredible. Someone comes back after a few days and edits the post. That is why editing time should be limited.

 

[Suzanne O]

What anti spam measures are you using?

These can be found in your admincp. Under user registration.

 

[RobinHood]

This is the type of crap that AI should excel at catching, and would be a brilliant use case. Doing insane moderation above and beyond what a simple filtering script or human moderator could catch.

Could be an interesting idea for a custom GPT to run each post submission and post edit through (or a local LLM on the server itself).

I absolutely hate when I have to contact mods or admins to edit or delete a post. So it would be user friendly in that you could leave editing (and deletion) permissions open to users, which is user friendly, but filter out dodgy edits using an AI mod.

 

[Sal Collaziano]

I also limit the amount of time a post can be edited after it was posted. But we can also set the initial registered user group to disallow edits at all, until they reach a certain amount of participation in the community - and then open up that ability to a new user group people are promoted to...

 

[FTL]

I despise these spammers. Thanks for the heads up.

 

[FTL]

It's not new. New members, post right after signing up. Nothing suspicious but later they went back and edit their posts to insert url. Sometimes using color code to hide the url.

Preventing post edits for something like the first 50 posts or so would help to discourage this.

 

[Sal Collaziano]

Preventing post edits for something like the first 50 posts or so would help to discourage this.

It certainly would...

 

[Alan_SP]

Why would you want to allow editing posts forever? On most content, you write something, see if it looks good, and move on.

 

[FTL]

Why would you want to allow editing posts forever? On most content, you write something, see if it looks good, and move on.

On my forum, I allow myself to edit posts forever and it's very useful. It does time out for my users though.

 

[AndyB]

adding link code in the quoted part of their responses. It's also setup so that it's not visible without viewing the code.

You might like this add-on:

https://xenforo.com/community/resources/remove-click-to-expand.8165/

 

[Sal Collaziano]

You might like this add-on:

https://xenforo.com/community/resources/remove-click-to-expand.8165/

That IS a nice add-on. As usual...

 

[Mendalla]

This is the type of crap that AI should excel at catching, and would be a brilliant use case.

But then the spammers would train their AI on how to avoid it and then antispam AI would learn how to catch that so the spammers AI would learn how to avoid that and the AIs would get smarter and smarter until ... BAM... Skynet.

I'll be back.