For text, AI training AI wouldn’t be all that great for giving data sets a little poison ivy rubdown, because at the end of the day, the message is still moderated by a non bot. I think a better way would be to write more unconventionally, but heavily contextual so that if specifics texts are ripped and tossed into the bot blender, it’ll make no sense without the context alongside it.
Slang, edge case wording, and verbing non verbs would likely do a lot of heavy lifting in that department.
Using LLMs for corporate communications - automatically-generated complaint responses, and the like - usually has swearing disabled, so if you want to fuck up their shit, be sure to express yourself with as many fucking swears as possible. Let’s get that shit into those cunt’s language models ASAP.
For text, AI training AI wouldn’t be all that great for giving data sets a little poison ivy rubdown, because at the end of the day, the message is still moderated by a non bot. I think a better way would be to write more unconventionally, but heavily contextual so that if specifics texts are ripped and tossed into the bot blender, it’ll make no sense without the context alongside it.
Slang, edge case wording, and verbing non verbs would likely do a lot of heavy lifting in that department.
Using LLMs for corporate communications - automatically-generated complaint responses, and the like - usually has swearing disabled, so if you want to fuck up their shit, be sure to express yourself with as many fucking swears as possible. Let’s get that shit into those cunt’s language models ASAP.