>>405 (OP)
One of OC's weaknesses is its strength: it's small and no matter how big it gets, it will never get to a size that allows mass spamming like 4chan without being immediately noticeable. For more subtle means, it can be detected and removed or moved ad-hoc. Personally, I believe the Indian posts in /hell/ are already a small-scaled attack by AI from our adversaries. If you observe the threads, they are nonsensical and pose open-ended questions which nobody responded to. I have some suspicion that the super negative posts that come out of nowhere in clusters from other flags aren't our friendly neighborhood doge, GNP spammer, or new anons that freshly arised from toxic environments such as 4trans outside of /hell/. I suspect either shills or bots similar to the bot that is discussed in your video.
Users have suggested setting up a reverse honey-pot by redirecting flagged glowniggers, shills, and others to it without their knowing for data mining. I suggest doing this. You've already acquired and tweaked anti-spam, anti-porn, and anti-opsec automation on this site. Such data can be given to knowledgeable anons, even if we can't use it ourselves to make a counter-shill AI tool in-house.
A rule of thumb that I derived from years of noticing is that shills and spam-bots may use swear words or slurs (kikes, niggers, saying jews are bad), but they do not post hard evidence of something (infographs of confirmed crimes by jews or glowniggers). Another advantage is that China and Russia really doesn't have a reason to against boards and places like this. Our adversaries are going to be glowniggers of the five eyes persuasion, independent kikes, willful or unknowing shabbos goy, paid poojeets, and israeli kikes; this gives us some insight of their psychology, their pain points, and what shape their agitprop will take the form of.
I partly made a thread suggesting cursive as a joke to filter anons, but something like this may actually be an interesting experiment as its own board down the road to see what happens when discussing topics that require a 99.99% purity test for genuine and intelligent anons.