LLMs enable at-scale deanonymization: up to 68% recall re-identifying HN and Reddit users
rickasaurus · x · 2026-09-25
A new arXiv paper by Simon Lermen, Nicholas Carlini, Florian Tramèr and colleagues shows LLM agents can deanonymize users at scale: with internet access, an agent re-identifies Hacker News and Anthropic Interviewer participants from pseudonymous profiles alone, matching hours of human investigator work. The pipeline uses LLMs to extract identity features, retrieve candidates via semantic embeddings, and verify matches. Unlike classic deanonymization work (e.g., Netflix Prize), it works on raw text across platforms. On three ground-truth datasets (HN–LinkedIn, Reddit movie communities, split Reddit histories) it substantially beats classical baselines, reaching up to 68% recall at 90% precision.
More from Safety
- Defending METR: public data can't second-guess its AI speedup findings — peterwildeford · 2026-11-11
- Australia's review of an OpenAI agent's unauthorized Medicare access exposes the agent accountability problem — cmb24k · 2026-11-11
- Adversarial RL envs for backdoor insertion vs. detection: where does it converge? — DimitrisPapail · 2026-11-11
- Using a third-party inference provider with your coding harness hands it shell access to your machine — TheZachMueller · 2026-11-11
- Democrats propose requiring federal permission to train AI models, with up to 20 years in prison — Polymarket · 2026-11-11
- WIRED asks 10 experts what the real AI nightmares are — not Skynet — AINowInstitute · 2026-11-11