Ok, so maybe it's not as serious as the post title proclaims but it's certainly becoming a "thing."
Yesterday there was an introductory post from a new member:
His profile picture showed a lovely image of he and his wife in hunting gear, with a setting sun in the background, as they kissed. Very appropriate for this type of forum. The picture had a watermark from a photographer in St Louis, which also seemed appropriate. But when I checked the users IP address it came up in Czechia: Not appropriate for this forum.
We get a fair number of these members and they seem to be getting more common. Some are obviously not real members, such as the ones @Gamma1214 enjoys playing with, and some are a bit more covert by making an on-topic reply about some piece of gear and then supplying a link where it can be purchased. But the trickiest ones are the ones that never have an obvious agenda. Some of them have been members for years, and made multiple posts that spawned good discussion and replies, before we noticed the pattern. I can only assume these are Ai performing some kind of market research as they generally ask for recommendations about a product (sleeping pads, mugs, etc).
I asked Chat GPT about the usage of Ai as forum members in general (yes, I get the irony) and it stated that along with the obvious purposes of pushing a product or agenda that they were also sometimes used for the relatively benign purpose of training the Ai to interact with people. This helps explain some of the more odd suspected Ai accounts, such as the one that would respond to posts with, "Yes, I see," once every few weeks.
I bring this up for a couple reasons:
1: Something to be aware of
2: If you suddenly find a post or poster missing this is probably why
3: I think it's really interesting and thought some of you might as well. If I wasn't a moderator who is supposed to watch out for this stuff I'd never be aware of most of these fake members. It's remarkable how good some of them are. The forum makes use of software that flags suspicious member applications and there are usually 10-30 of these per day (it's been as high as 90/day) but some still slip through the filter.
Alan
Yesterday there was an introductory post from a new member:
His profile picture showed a lovely image of he and his wife in hunting gear, with a setting sun in the background, as they kissed. Very appropriate for this type of forum. The picture had a watermark from a photographer in St Louis, which also seemed appropriate. But when I checked the users IP address it came up in Czechia: Not appropriate for this forum.
We get a fair number of these members and they seem to be getting more common. Some are obviously not real members, such as the ones @Gamma1214 enjoys playing with, and some are a bit more covert by making an on-topic reply about some piece of gear and then supplying a link where it can be purchased. But the trickiest ones are the ones that never have an obvious agenda. Some of them have been members for years, and made multiple posts that spawned good discussion and replies, before we noticed the pattern. I can only assume these are Ai performing some kind of market research as they generally ask for recommendations about a product (sleeping pads, mugs, etc).
I asked Chat GPT about the usage of Ai as forum members in general (yes, I get the irony) and it stated that along with the obvious purposes of pushing a product or agenda that they were also sometimes used for the relatively benign purpose of training the Ai to interact with people. This helps explain some of the more odd suspected Ai accounts, such as the one that would respond to posts with, "Yes, I see," once every few weeks.
I bring this up for a couple reasons:
1: Something to be aware of
2: If you suddenly find a post or poster missing this is probably why
3: I think it's really interesting and thought some of you might as well. If I wasn't a moderator who is supposed to watch out for this stuff I'd never be aware of most of these fake members. It's remarkable how good some of them are. The forum makes use of software that flags suspicious member applications and there are usually 10-30 of these per day (it's been as high as 90/day) but some still slip through the filter.
Alan
