"Hit or miss" is the perfect description for me as well.
Below was 5/1 and this morning , 5/5...5 mins later, back up again. No telling how many times the last several days it's the same scenario, I don't camp out on this keyboard.
Lately a lot of small forums are experiencing slow downs and hangs. Such forums are being hit by 1000's of AI-searchbots continuously for days.
Larger forums have the means to maintain a staff that can block these "attacks." Little forums, especially with no one at the rudder, just get hammered into the ground.
Great, this could lead to a lot more bot posts here. A much smaller forum I am a moderator for is set that the first post from a new member must be approved before appearing in the forum. If the new user looks legit, his status is changed to unmoderated.
Great, this could lead to a lot more bot posts here. A much smaller forum I am a moderator for is set that the first post from a new member must be approved before appearing in the forum. If the new user looks legit, his status is changed to unmoderated.
Keep in mind the latest AI generated posts now look legit. Sometimes these automated bots post interesting questions, other times they provide assistance that seems valid. All the while, we humans are unaware that they aren't real people that need real help. None of this "helpful" activity is done to assist us.
Along with a lot of other reasons, these fake interactions allow them to improve their language models. Because the people operating these things have a long term goal that is not in our best interest. Letting them get away with it just means they get increasingly better at fooling us every day.
Unfortunately it is getting harder to detect them. I've heard that the majority of social media posts are now either human assisted AI, or fully automated AI. And now these AI actors are spreading to every forum they can get into.
I'm not AI. Or maybe I am. It's getting hard to know for sure.
I think it’s a bit over blown as to the prevalence of AI posts, certainly here. They’re fairly easy to spot
As to broader social media, I dunno. I participate on X a lot but really don’t see many. Some complain about them but I’d say 90%+ are in relation to DMs and otherwise silent friend type requests as opposed to conversational posts in a thread.
I disagree. They are getting very convincing. Sometimes impossible to tell without a time consuming investigation on the web history of a new forum poster.
Most of the bot posts are well written, to the point where I'm impressed. What makes them easy to spot is context. Usually they're the first post from a new member, combined with replying to an old thread, pushing a product, or going off with too many technical details.
I think a lot of the initial posts are fake. They're too detailed, too perfect, too long, and too unusual. And the first post of a new poster on a very old topic is a red flag.