dorsal/arxiv
View SchemaAgentic LLMs as Powerful Deanonymizers: Re-identification of Participants in the Anthropic Interviewer Dataset
| Authors | Tianshi Li |
|---|---|
| Categories | |
| ArXiv ID | 2601.05918vv1 |
| URL | https://arxiv.org/abs/2601.05918 |
| License | http://creativecommons.org/licenses/by-sa/4.0/ |
Abstract
On December 4, 2025, Anthropic released Anthropic Interviewer, an AI tool for running qualitative interviews at scale, along with a public dataset of 1,250 interviews with professionals, including 125 scientists, about their use of AI for research. Focusing on the scientist subset, I show that widely available LLMs with web search and agentic capabilities can link six out of twenty-four interviews to specific scientific works, recovering associated authors and, in some cases, uniquely identifying the interviewees. My contribution is to show that modern LLM-based agents make such re-identification attacks easy and low-effort: off-the-shelf tools can, with a few natural-language prompts, search the web, cross-reference details, and propose likely matches, effectively lowering the technical barrier. Existing safeguards can be bypassed by breaking down the re-identification into benign tasks. I outline the attack at a high level, discuss implications for releasing rich qualitative data in the age of LLM agents, and propose mitigation recommendations and open problems. I have notified Anthropic of my findings.
{
"annotation_id": "9eb52b9b-64e2-415b-beed-c7711e3c8ac2",
"date_created": "2026-02-17T05:53:05.106000Z",
"date_modified": "2026-02-17T05:53:05.106000Z",
"file_hash": "2fe70c329f304276474aaf1b93fb17504c6f308b43e0ed911e6e58c02ba7a12e",
"private": false,
"record": {
"abstract": "On December 4, 2025, Anthropic released Anthropic Interviewer, an AI tool for running qualitative interviews at scale, along with a public dataset of 1,250 interviews with professionals, including 125 scientists, about their use of AI for research. Focusing on the scientist subset, I show that widely available LLMs with web search and agentic capabilities can link six out of twenty-four interviews to specific scientific works, recovering associated authors and, in some cases, uniquely identifying the interviewees. My contribution is to show that modern LLM-based agents make such re-identification attacks easy and low-effort: off-the-shelf tools can, with a few natural-language prompts, search the web, cross-reference details, and propose likely matches, effectively lowering the technical barrier. Existing safeguards can be bypassed by breaking down the re-identification into benign tasks. I outline the attack at a high level, discuss implications for releasing rich qualitative data in the age of LLM agents, and propose mitigation recommendations and open problems. I have notified Anthropic of my findings.",
"arxiv_id": "2601.05918",
"authors": [
"Tianshi Li"
],
"categories": [
"cs.CR",
"cs.AI",
"cs.CY"
],
"license": "http://creativecommons.org/licenses/by-sa/4.0/",
"title": "Agentic LLMs as Powerful Deanonymizers: Re-identification of Participants in the Anthropic Interviewer Dataset",
"url": "https://arxiv.org/abs/2601.05918",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "aacf6f37-6ba5-48dd-94d9-f89b573c2808",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}