Its basically a cost question.
There's a bit of a time lag of course but in general as long as the benefits (influence, money (via scams, advertisement, training data creation, ect)) to running the AI on B12 outweigh the costs we'll get a bunch of AI posters. If its questionable then we'll only get a few, and if its pretty clearly not worth it then we won't get any.
When compared to traditional social media the benefits are small for influence in that only a small number of people will see the posts, and small for money in that its much harder to do scams and advertisements here then on traditional social media.
Similarly the cost is fairly high, a post on traditional social media requires very little context. Typically it only needs to read an initial post/image, and respond with a few lines of text. On B12 you want to keep an entire page or even multiple pages in context. In addition the chance of being detected and banned if you actually try to do anything is much higher then on sites like imgur or reddit because of the small size and current issues in AI (in this regard the cost will go down exponentially with increases in AI capabilities).
---
I can't even begin to predict how much money an AI would make per word on B12, but given the costs I don't think we are anywhere near it yet.
Of course "it being 200x too costly", and "oh, yeah, AI got 50x cheaper and 10x better at getting money out of people" is the type of gap that AI at current rates of advancement can easily cross in 2-3 years, so *shrugs* its really a big who knows from me.
---
In actual AI news:
https://thelibre.news/foss-infrastructure-is-under-attack-by-ai-companies/Web scrapers for AI companies massively screwing over websites by spamming them with requests and repeatedly downloading everything over and over has become a real problem, one without a real solution.
https://arstechnica.com/ai/2025/03/cloudflare-turns-ai-against-itself-with-endless-maze-of-irrelevant-facts/"When we detect unauthorized crawling, rather than blocking the request, we will link to a series of AI-generated pages that are convincing enough to entice a crawler to traverse them," writes Cloudflare. "But while real looking, this content is not actually the content of the site we are protecting, so the crawler wastes time and resources."
Cloudflare at least now has a solution, creating a ton of fake AI created web pages to screw with these crawlers, making them waste their time and making the training data they collect worthless, although of course we'll see how well it ends up working.