Question: I’m wondering what to do the next time I spot a suspicious post. What should we do when we see a post that appears to come from a bot and/or generative AI? Flag the suspicious post, or is something else better?
Sometimes I read HO before my coffee kicks in, so I want to be sure I’m doing the preferred thing. 
When in doubt (well, not really 
) I try to be courteous and ask the poster a question or two. Usually, the same AI garbage comes as a reply. Then I flag & have the mods take care of it.
I just flag it.
Either of what @linguafood or @small_h said. Sometimes it’s VERY easy to figure out, based on wording, that it’s a ChatBot or spam, sometimes not. But report and let the Mods review.
The generative AI accounts lately seem to use the same username and photo across multiple websites, and their content posted in those other websites is often the same general drivel you see here that arouses your Spider sense.
I PM’d to ask one of the mods about suspected bots and the response I got was it’s better to give the benefit of the doubt, take a wait-see approach.
So rather than flagging (which I believe in this iteration of Discourse might have an automatic effect of graying out the post), I’ll sometimes send a note to the mods with a link to the suspect post.
More often I’ll do what I’ve seen Natascha (@linguafood) do, as she mentioned above, which is just ask a nice question to see what results.
Just flagged one.
I can usually sense AI content from a mile off. It’s literally regurgitated and overly dramatic especially when it comes to things like food. I’ve never heard people use terms like “symphony”, “elevate”, etc to describe food / eating until this past year!
Hmmm, what about new poster(s) who show up, post a question, and then don’t engage further while the HO community answers? My spidey sense is that’s a bot.
I have read that AI bots are invading forums like ours to experiment. Ugh.
Or maybe they just don’t like us?
Or at least me.
Well, then they don’t seem to like me either.
No response to a gentle question that I posted on one of that new user’s topics, which I had hoped would draw a response. And help to establish they are human rather than a bot.
They tend to out themselves eventually. If yer iffy about a poster you can flag them for the mods (although they tend to be on top of this shit in general). I usually give them the benefit of the doubt initially, but if I find them sus I just don’t engage any further 
I believe I know the new user in question. Odd thing is, unlike the typical bot, this user has read other threads. Not to mention, the posts have generated some interesting discussion. So I too am inclined to extend the benefit of the doubt for now,
Sigh . . .spam is so much easier to spot and squash than AI bots. And I wouldn’t want to discourage new users without giving them a chance.
Thanks, @gaffk! Your description maps to the behavior that is giving me pause, so I do believe you have noticed the same thing. Sounds like we should wait to see where things go.
I flagged it, then found this discussion.
For what it’s worth: When I can see that a new user doesn’t include personal perspective, cites only commonly known information, and doesn’t engage with our questions that poster is likely an AI bot. Also when the same new user name and avatar show up as new on other forums on the same date they join HO.
We’re a helpful community of humans. It saddens me when invading bots waste our time. And possibly use the content we create to feed the AI’s learning. Sigh.
Yep. But at least this latest one started an interesting conversation before losing the benefit of the doubt. I will say AI is improving.
I want to clarify the impact AI has on our community.
There have been cases where users, mostly new, posting texts generated by Generative AI engines such as ChatGPT. Usually those posts are generic posts without useful information. We just remove those since if you want to talk to AI you can go to ChatGPT, etc. No need to come to a forum where you are expected to interact with fellow humans.
AI hasn’t advanced to a point where it can just sign up for a new account, validate and start navigating the site and posting. It’ll happen at some point in the future, probably when AGI (artificial general intelligence) matures. but right now generative AI trains itself with a large dataset and tries to make sense of patterns and logic. The more specific the job we ask Gen AI to do the better they do.
So it isn’t like bots posting AI generated content on the site. its humans posting AI generated content.
There are, or should I say, were, AI bots crawling the site. These bots operate similarly as the search engine bots. For example, Google bots crawl our site and index our pages so you can find the info from our site on Google. ChatGPT bots crawl our site presumably to train their data set so their Gen AI engine can get better. So far, I only see ChatGPT bot crawling the site, and I requested them to stop their crawling. Legitimate companies usually honor the ‘stop crawling’ requests. I haven’t seen the ChatGPT bot come back after the request.
Thank you for taking the time to give us some background; that was certainly enlightening.
Hopefully, the bot activity will cease.
Thinking more about that one - its first few posts had some of the characteristics of the AI (overly effusive) but still had me wondering. That “help I think I’m overloading my pizzas” thread (paraphrased as it’s now gone) could have come from any of us.
But that final straw, where it lost the benefit of the doubt, as you put it - I wonder if that was intentional? The other posts were at least quasi-original material, but in that last one it just straight out plagiarized the HO member who started that thread.
Kind of like, “Hey, if you idjits ain’t figured me out yet, I’ll get brazen about it”.
ETA
Thanks, didn’t know how it worked.
default bot behavior. when it can’t generate something, just quotes the old stuff… because obviously no one will notice . . .
there are a number of things that tip off spammers and bots. altho a lot of people use VPNs and proxy servers, when the IP bounces all over the world for someone stating "I’m in . . " it’s a red flag.
the stilted language and unusual word usage is another flag - real people don’t talk like that . . .
and when a user makes 4-20 post at warp speed (i.e. post time - sometimes 2 or 3 in the same minute - ) . . . that’s not a human changing sections and reading posts.
