Safety
153d agoLLMs Enable Large-Scale Online Deanonymization
Large‑scale deanonymization using LLMs threatens online privacy worldwide. Researchers from OpenAI and Google demonstrate that models can infer personal details from public posts, linking pseudonymous accounts to real identities. This capability raises concerns for social media platforms, governments, and civil‑society groups.
The Signal
The technique underscores the need for stronger privacy safeguards and regulatory oversight across the global digital ecosystem.