"Language models like Gemini and ChatGPT did in minutes what would take a human several hours, if they could even manage to crack it at all. The models identified 68% of anonymous users with 90% precision compared to near 0% for the best non-LLM method, according to the scientific article."
"People sometimes express their opinions through pseudonymous accounts, assuming that those opinions will remain private. The existence of a mechanism to investigate or monitor with large language models that allows us to simply ask about a person's beliefs, political opinions, insecurities, or anything else that can be extracted from their anonymous Reddit account could disempower many people today."
"Our results show that the practical obscurity protecting pseudonymous users online no longer holds and that threat models for online privacy need to be reconsidered. It's not even necessary to dox a person to affect their behavior: AI can already reveal a lot of personal information of pseudonymous accounts on message boards and social media platforms."
Researchers demonstrated that AI language models like Gemini and ChatGPT can identify anonymous social media users far more effectively than traditional methods. By analyzing thousands of posts from platforms like Reddit and Hacker News, the models achieved 68% identification accuracy with 90% precision, compared to near 0% for non-LLM approaches. This capability threatens the practical obscurity that previously protected pseudonymous users online. The implications extend beyond simple de-anonymization; AI can extract personal beliefs, political opinions, and insecurities from anonymous accounts, potentially disempowering users who relied on anonymity for free expression. This technological shift has prompted legal disputes, including between Anthropic and the Pentagon regarding government use of AI for de-anonymization purposes.
Read at english.elpais.com
Unable to calculate read time
Collection
[
|
...
]