What's new
What's new

Board policy on chat bots and AI?

Donkey Hotey

Stainless
Joined
Dec 22, 2007
I was going to PM a mod this question but, figured the wider audience needs to know and/or might want to weigh in with their views. The question is: what is the future direction of the forum with regard to chat bots and AI content? Does it get left in the threads if it isn't spam and it's technically correct? Is this forum going to become a repository of Quora-like answers created by different bots chattering at one another?

In the past 24 hours, I've seen two instances of content that was painfully obvious as AI content. As the bots infiltrate the place, is content that isn't advertising something, still considered junk and the mods want to know? I don't want to flag it if it's being allowed in intentionally.

I'm 99% this is a bot, including an AI generated profile picture to go with it. If the picture isn't AI-generated, they got it from a Blaser brochure.


Here's another. It reads like a trifold brochure from the paint store. It's not selling anything but, a person could go to a paint website if they wanted this kind of answer.


Personally, I don't want to hear from a Sandvik chat bot either. If I post a question, I'd like the posts to be from living, breathing humans who have actually used or done whatever. I need someone that has real experience, not the results of a bot that scraped the internet. But it's not my forum and not my strategy. What do the mods want done with this content in the future? Do we report it?
 
If I'm completely sure, I'll delete them. Like I just did with the DIY Paint user. It's not always easy to tell though, some people do just write oddly... If I'm not positive, I'm going to err on the side of caution. The paint one was easy; nobody in their right mind is going to sign up here just to post about paint. The first guy has been registered here since 2021, little trickier. He may be legit.
 
No, I was talking about post #10 on the chat bot thread. I don't even want to post the obvious ways it's a bot because it will scrape this information and improve next time.
 
Remember that many here do not speak English as a first language. They may be using some kind of translation bot both ways. I am sure that makes their language seem odd and bot like.
Interesting to read UK and Australian shop slang. I started out reading my dad's old model machinist magazines he keep in the basement above the shop table.
Anyone can claim to be from anywhere in the world to excuse their poor bot english. It is interesting to see folks from Africa and Russia posting here. I jump on folks asking for advice about shop heating/cooling, lighting with no mention of climate or location.
Bill D
 
The reverse image search on that user's picture comes up as a supposedly legit engineer, employed in the role of machine learning. That doesn't mean that it's not a made up person. Or maybe he's legit but, turned his bot loose on the forum as a playground to improve the learning.
 
I answered a strangely worded post on here about a week ago by quoting an AI generated answer. I didn’t do it to be rude, but rather just to pay back the post in kind.

This kind of posting is an extension of the thinking that if it’s on the Internet, it’s free to copy and use any way you want to. They are essentially mining other peoples intellectual ability to teach their machine for free.

while it’s intellectually interesting, I think the biggest problem is that without natural consequences the machine learns and escapes supervision. So there’s no downside to being wrong or right. In my world, the creator or creators and owners of AI machines should be held strictly liable for the results that are output and any harm that comes from that.
 
If its an obvious bot, I'll nuke it as well. However it's not always obvious if its bot, if it's someone who has been too many MBA classes or if its someone who is using a translator. #10 in the chat bot thread probably is a bot, but its just a single post, and it doesn't have any links or any other nefarious/inappropriate content. It's better to just ignore a single post like than rather than nuke it and risk alienating a new user who may just be trying to fit in. Now if its got a link in it, or if the user keeps posting (or if a flood of similar posts show up) It's a different story.

I'm not concerned about someone using the forum to train their bot, they aren't going to get any thing useful out of making a single post every once in awhile. If they are reading all the data in the forum, thats a bigger question, beyond the scope of any of us at the mod level.
 
No, I was talking about post #10 on the chat bot thread. I don't even want to post the obvious ways it's a bot because it will scrape this information and improve next time.

The post #10 probably is a spammer or bot, but there's no positive way to say for sure. I usually just look at post content and history before making a call on those. If they continue posting repeated messages that look like they come from a catalog text or if they post a spam link, it's adios. If there's technically no violation of the rules, they are safe. For the moment.
 
Just wait until the AI bots ingest thousands of Thermite posts. Because then it will be too late.

"It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear! And it absolutely will not stop, ever, until you are dead!"

Oh wait, that's the Terminator.
 
Just wait until the AI bots ingest thousands of Thermite posts. Because then it will be too late.

"It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear! And it absolutely will not stop, ever, until you are dead!"

Oh wait, that's the Terminator.
I thought Thermite was a bot.

Regards Tyrone
 
If there's technically no violation of the rules, they are safe. For the moment.
I spot others from time to time. I look them over and don't report unless I'm absolutely sure as well. Some are weeks or months old accounts and--yeah--they seemed to be planted to avoid calling immediate attention.

How many new sign-ups do you get per day here? I'm a mod on another board. We only get a few a month since it's a pretty obscure place. I usually catch the spammers within an hour or two but, that's because i have the luxury of getting notifications of every new account and there just aren't many.
 
I am a moderator on a small forum. If I at all suspect that a post or poster is spam/bot/AI it’s buh bye. I tend to shoot first and ask questions later. I do check IP origins as best I can. I do not ever post how or why I think posts are bogus. Let the idiots behind that garbage keep guessing.
 
We don't get access to that info. I've accidentally canned one or two users that I thought were spammers that weren't, but I try to keep a close eye on signup date, post history, IP info, etc. before I shitcan anyone. Don't want to create too many headaches for upstairs...
I get an email as soon as a new account signs on. Wasn't sure how many other forums use that setting. It might be more than you want to deal with on here.
 
We don't get access to that info. I've accidentally canned one or two users that I thought were spammers that weren't, but I try to keep a close eye on signup date, post history, IP info, etc. before I shitcan anyone. Don't want to create too many headaches for upstairs...
I think we need to sick AI on identifying the AI
Let them talk to each other for a while and see what happens
 
I get an email as soon as a new account signs on. Wasn't sure how many other forums use that setting. It might be more than you want to deal with on here.
We get about 20-30 new signups a day, not sure how many are actual users vs bots though.
 








 
Back
Top