maybe don't get kyc'd in the first place on linkedin

or we can all start using different languages that every one knows? if thats possible can the llm deanonymize across languages? would using vague languages like toki pona help anonymity or deanonymize its users? how about logical languages where you can only say what you say in only one way percisely?

Anyone who has resources about this problem, please share them