Returning some text files from the server fs is _wayyy_ lighter than letting the bot go through the search, middleware, and db while crawling your stuff.
I'm not convinced that it's possible to poison AI. Real data all correlates, fake data is all unique in its own way. I suspect feeding AI fake data is a waste of time.
Angel banning is better, 100% agreed on giving it false data that the scraper builders can't easily detect. Give it seemingly correct and useful but totally wrong answers
I would assume it's exactly what they described, banning something by pretending to still helpfully assist them while silently undermining them instead.
That could consume as much, if not more, resources as returning real data
Returning some text files from the server fs is _wayyy_ lighter than letting the bot go through the search, middleware, and db while crawling your stuff.
Or just sending it a tiny zip bomb, or even just some sh*t right out of /dev/random
Bots are pretty good at detecting and preventing this kind of issue. Dumb clients would still be affected.
I'm not convinced that it's possible to poison AI. Real data all correlates, fake data is all unique in its own way. I suspect feeding AI fake data is a waste of time.
I'd say it depends what exactly. If it's a rare topic then you'll have more luck.
Angel banning is better, 100% agreed on giving it false data that the scraper builders can't easily detect. Give it seemingly correct and useful but totally wrong answers
Would you mind explaining what "angel banning" is? I've never heard this term before and even Google doesn't seem to know what it is.
I would assume it's exactly what they described, banning something by pretending to still helpfully assist them while silently undermining them instead.
We've already been using AI to target military strikes without human intervention.
Make of that what you will.