Nytro Posted February 27 Report Posted February 27 We show that large language models can be used to perform at-scale deanonymization. With full Internet access, our agent can re-identify Hacker News users and Anthropic Interviewer participants at high precision, given pseudonymous online profiles and conversations alone, matching what would take hours for a dedicated human investigator. We then design attacks for the closed-world setting. Given two databases of pseudonymous individuals, each containing unstructured text written by or about that individual, we implement a scalable attack pipeline that uses LLMs to: (1) extract identityrelevant features, (2) search for candidate matches via semantic embeddings, and (3) reason over top candidates to verify matches and reduce false positives. Compared to classical deanonymization work (e.g., on the Netflix prize) that required structured data , our approach works directly on raw user content across arbitrary platforms. We construct three datasets with known ground-truth data to evaluate our attacks. The first links Hacker News to LinkedIn profiles, using crossplatform references that appear in the profiles. Our second dataset matches users across Reddit movie discussion communities; and the third splits a single user’s Reddit history in time to create two pseudonymous profiles to be matched. In each setting, LLM-based methods substantially outperform classical baselines, achieving up to 68% recall at 90% precision compared to near 0% for the best non-LLM method. 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. Download: https://arxiv.org/pdf/2602.16800 Quote
fbi_suge Posted February 27 Report Posted February 27 Ideea e interesanta insa sa combati analiza e f usor: 1. Nu dai detalii despre tine 2. Dai detalii fake despre tine 3. Folosesti software de parafrazare Quote
Nytro Posted February 27 Author Report Posted February 27 Da, nu e tocmai practic research-ul lor, dar e destul de interesant ca metodologie. Ideea de baza, desigur, e sa nu dai detalii despre tine niciunde. Degeaba esti "HackerMan1337" daca ai Facebook-ul la fel. 19 minutes ago, fbi_suge said: 3. Folosesti software de parafrazare Da. Sau LLM-uri, doar sunt bune la asta. Quote
fbi_suge Posted February 27 Report Posted February 27 (edited) as adauga 4. Trb sa iei in considerare din prima zi ca gaborii americani iti stiu identitatea reala de la bun inceput. Exista vreo asociere reala intre contul tau bancar (de Persoana Fizica) si conturile alea bancare? E vreo asociere? toate soft-urile de parafrazare folosesc modele de AI. Edited February 27 by fbi_suge Quote
