The real answer is just moderators, ideally with a range of lifestyles and forum obsession to cover the late night / early hours spamming that is more likely to get through
Anti bot sign up methods all get beaten somehow. Many questions have the answer in the question, therefore a bit of AI can often identify the answer or at least learn from failures. If the forum locks out x number of fails then the bot just needs to report back it’s failures for another to try, or put the question to a human and then the answer can be shared around
Captcha Am I a robot text images worked for a while until AI could learn to read them (Bearing in mind back in the late 80s/early 90s as a student I did a site visit to see an international distributors handwriting recognition system, which they didn’t call AI but the system was learning to read & guess hand written forms with two clerks each confirming / correcting every guess
The picture Captchas aren’t really checking that you get the answer right, but measure your mouse movements. A bot would have been expected to be efficient and a human to be wobbly. But AI bots can learn to act like a human with ‘randomised’ hesitation and wobble
(But don’t worry too much about AI taking over the world, ask ChatGPT who the prime minister is, correct their answer and see what response you get)
Post filtering, automated report handling etc are a pain in the arse. They will probably harass more human members then bots, and become a denial of service attack on a forum member that someone doesn’t like.
Some of these can be configured into forum software fairly easily and others are likely to be much more effort than some semi willing voluntolds
The recent spam may look like a lot and is frustrating for the viewer. Moderators and the software could be successfully smiting more before it gets through…… On one forum a few years ago went through a couple of days of constant denial of service spam bot attacks, and at times we were blocking each other out of the forum with IP range blocking ….. a lot of SMS & emails were circulating to find a moderator who could still get in and green light friendly ranges to allow the rest of us back in