I got hit by a word filter too. Each individual word was SFW but the combination could be interpreted as hateful (even though it wasn’t in context). So they have some sort of intent-matching system that is poorly implemented, probably piping it through the cheapest haiku model they can access.
this has been happening for some time already. bans after saying a wrong word happen immediately within a few seconds
not evenposting, even “looking” at your browser wrong is enough. if the AI assumes your account is botting or something it will ban it.
I got hit by a word filter too. Each individual word was SFW but the combination could be interpreted as hateful (even though it wasn’t in context). So they have some sort of intent-matching system that is poorly implemented, probably piping it through the cheapest haiku model they can access.