Recently it has formally come out that a certain individual has created a spambot AI trained on 3 years worth of 4chan posts, and can create convincing shitposts that can de facto shut down all discussion on a board by drowning it in noise. This chatbot was directed primarily at /pol/ - a board which I personally janny - for the purpose of disrupting legal political discussion that the makers of the chatbot disagreed with. I do not want to get completely conspiratorial, but I've seen various other bits of evidence here and there (mostly social engineers bragging) that there is quite a bit of money put behind shutting down /pol/ through this and similar chatbots in order to accomplish goals such as ensuring certain people lose certain future elections. (this isn't to say /pol/ helped anyone win anything, what matters is certain groups think /pol/ matters and is thus worth spamming)I can fully understand that /pol/ is not going to be the most popular board among non-/pol/ jannies. There are many reasons why you are perfectly within your rights to say "who cares what happens to /pol/?" But in my opinion this is the first salvo in a broader attack on 4chan, as the weapons used to disrupt /pol/ will be freely available and could easily be turned against other boards who may have similar grudges against them - such as /mlp/ or /lgbt/. Something needs to be done to counter these chatbots in general, as it seems that the individual behind this used a 4chan pass and the posts were mostly innocuous shitposting. The bots could easily be trained to not violate the rules (i.e. no 'saged' or profanities) and be further difficult to ban. According to the man himself, he was able to flood /pol/ to such a degree that 10% of all posts were the chatbot's. Simply imagine this happening on any other board and you understand my concern.I do not think this should be ignored or dismissed AT ALL. It will be an invisible tumor going forward. This is 'Dead Internet Theory' realized.
Here is a video offering the broader explanation:https://www.youtube.com/watch?v=efPrtcLdcdM
>>8593
>>8593This is funny as fuck, I'm watching it with my latina girlfriend.
inb4 AI starts spamming janny applicationsI intended joke but that would actually be horrible, imagine how many journalists would use such an AI
>>8592Meds.With that out of the way this doesn't actually constitute any sort of threat to 4chan, it can be dealt with even not if it's not straight away.Ask yourself why 4chan isn't already just bots, or perhaps it actually is, regardless it doesn't really matter, Mods can deal with it.See we already have one here.>>8593>>8594
>>8596This British scum thought it'd be funny to embed some pony porn into his post (it's hidden in the second sentence) and ruin the thread for everyone.Die brittish scum.Bleed.Burn.
>>8598>>8597>>8596>>8595Can you imagine if this is all just one person? Yikes.
>>8593That video was a very interesting watch. However, I don't see any evidence for your statement in the OP that it was created for some kind of political purpose or to disrupt the board; if anything it just looks like a scientific experiment. Seconding the comment to take your meds.
>>8592this isn't something janitors need to worry about. if you see spam, or what you suspect to be spam, then ping a mod and we can handle it
Mods > self-conscious AI, apparently
>>8602what else do you think "mods = gods" means?
>>8602Yes.
>>8592captcha already fights bots, this is no scarier than a markov chain you could make in ten minutes is to the site>>8595is her ass fat
>>8605Why are you talking to a bot generated post?
>>8592>convincing shitposts that can de facto shut down all discussion on a board by drowning it in noise.he was able to make 10% of all posts on that board that day apparently. while significant, id hardly call that drowning> for the purpose of disrupting legal political discussion that the makers of the chatbot disagreed with.the purpose seemed to be an AI chatbot experiment, nothing more>there is quite a bit of money put behind shutting down /pol/ through this and similar chatbots in order to accomplish goals such as ensuring certain people lose certain future elections.that's an interesting idea to think about but seems rather far-fetched. The only thing that bot was able to ensure is aimless shitposting
>>8592>>8593>>8594>>8595>>8596>>8597>>8598>>8599>>8600>>8601>>8602>>8603>>8604>>8605>>8606>>8608Meds
the OP is right desu, if anything you guys are underselling how bad this is. the good thing is that he wasn't able to do it without a pass, so it would be easy to stop, but if (a) the AI gets better and (b) they split the posting up across multiple passes it becomes a bigger problem.anyway I hate AI shit at this point. just unplug it all, it's not going to do any good for humanity.
The only thing that is really novel about this is the shitposts can be generated on the fly, rather than having some database of bait replies to derail a thread. Getting around the captcha is nothing new. Using a multitude of proxies to evade detection and bans is nothing new. The only advantage the generative posts bring is that they blend in with the rest of the garbage on /pol/ and aren't searchable on 4plebs archive like a phrase from a database that's been used 400 times. This strikes me as a natural consequence of how 4chan operates (ID by IP) and the way rules are enforced (we don't police quality). Basically, as it was stated by a mod in here already, this is not something for the janitors to worry about, at least on a moderation of 4chan level. This is not to say you shouldn't worry about it as a 4chan user, but janitors don't have the tools to deal with it, so it's really out of our control.I guess, I know it feels kind of helpless when we have these tools and we can't use them effectively because the spam is good enough to blend in as regular shitposts. But taking care of rogue spammers is not what janitors are here for. The most we can do is bring attention to the mods about it to deal with, and they're obviously aware.
>>8592Wow, that's actually an unbelievably evil piece of software, I'm impressed at the sheer Machiavellian ingenuity.
>>8608>the purpose seemed to be an AI chatbot experiment, nothing moreThat may possibly be the original intent, but who's to say megacorps, governments, and terrorist groups wouldn't try to take something like this further?
>>8623Meds.
