dailyai.report

Large‑scale deanonymization using OpenAI LLMs shows that even sparse user data can reveal identities, exposing privacy risks across social media, e‑commerce, and public forums. Researchers demonstrate that language models infer demographic and behavioral cues from text, enabling targeted profiling.