AI or human? Any smart ways to tell?

Alvin63 said:
If you try it, after installing you need to turn it on in permissions for whichever user group you want to see the results (I just have it on for admin) and the link to check what it's doing shows up in the sub-links under Forums :-)
Click to expand...

Basically it should be turned on for all registered? Or are you sayhing turn it on for guests too?
 
creativeforge said:
Basically it should be turned on for all registered? Or are you sayhing turn it on for guests too?
Click to expand...
It's on automatically in options after installing - but for the statistics I just have the permission on for admin so I can see the results - I didn't really think members needed to see the results of what was being blocked.

It just works to stop them registering. If you have any that are already registered you'd just need to work out which ones they are and spam clean them I guess.
 
As the AI models compete against each other, they will attempt to poison each other by provided automated false information so that the other models start making mistakes.
 
I think they will fall for it and just read the code/see the invisible fields. I think it would be an unusually expensive option for a botmaster, for an AI bot to include a scanner to read a page (I might be wrong!)
 
Alvin63 said:
It's on automatically in options after installing - but for the statistics I just have the permission on for admin so I can see the results - I didn't really think members needed to see the results of what was being blocked.

It just works to stop them registering. If you have any that are already registered you'd just need to work out which ones they are and spam clean them I guess.
Click to expand...

I certainly do not want everyone to know about this. I want it to run secretly in the background. But I'm wondering now if they succeed in registering before, it means they are good enough to create emails and respond to them to fool us. However if this leads into oblivion, that;s what I want.
 
That's what it does - runs secretly in the background. The permission is just to be able to see analytics and results so you could just turn on the permission for yourself or only admins/moderators. I think I'd be concerned about the ones on site currently sending dm's to people! If you have dM's on. If you know which ones are suspect maybe turn off DM's for them individually and set them to manual approval for now? Can't quite work out how they would send emails - they might but presumably that would be spotted as spam by the email provider.
 
OK, so nobody but those with the permission can see it? I was taken aback, it's not something I wanted to have openly, I would prefer to access via the admin panel. :)
 
No that link only shows for anyone that has the permission turned on. Can you see loads of bots being stopped? I eventually turned off the permission for myself and just left it to do its thing as I was getting obsessed watching it all the time 😂
 
Alvin63 said:
No that link only shows for anyone that has the permission turned on. Can you see loads of bots being stopped? I eventually turned off the permission for myself and just left it to do its thing as I was getting obsessed watching it all the time 😂
Click to expand...
Yeah that's what I was afraid of. "DID YOU FORGET TO CHECK THE LOG TODAY?? GO NOW!! DON'T WAIT!!! ONE MORE TIME CAN'T HURT!!!"
I'll leave it there for a few days just to make sure the other admins are also aware.
Thanks for all the help! (y)
 
Mendalla said:
As I understand it a lot of universities and school systems subscribe to software (AI driven itself, ironically) that they can run papers and such through. Whether that's cost effective for a forum site is probably questionable.

Going back to in person exams and banning use of personal devices in the exam room is something else that I have heard done but that doesn't help us.
Click to expand...
i think it could be cost effective for a forum site....really i think all we need is a note above the sig line like "AI content used in this post" or whatever....then if its abused too much people can be addressed in private and if they dont reply and persist etc the solution is obvious and victimless.

we already use AI for tag creation so if its already scanning the content it just needs one more task added on that operation, so someone needs to make an addon for that at least :whistle:
 
You must log in or register to reply here.
Back
Top Bottom