This capability comes for free…
Large-scale online deanonymization with LLMs
ArXiv has an interesting paper on mechanising online deanonymization attacks on large datasets. Even if you’re not interested in LLMs it’s still worth understanding what they’re capable of and how they may be deployed against you in the future.
The ArXiv page can be found here and the paper’s PDF can be found here.