Research

Paper

AI LLM March 10, 2026

CLIOPATRA: Extracting Private Information from LLM Insights

Authors

Meenatchi Sundaram Muthu Selva Annamalai, Emiliano De Cristofaro, Peter Kairouz

Abstract

As AI assistants become widely used, privacy-aware platforms like Anthropic's Clio have been introduced to generate insights from real-world AI use. Clio's privacy protections rely on layering multiple heuristic techniques together, including PII redaction, clustering, filtering, and LLM-based privacy auditing. In this paper, we put these claims to the test by presenting CLIOPATRA, the first privacy attack against "privacy-preserving" LLM insight systems. The attack involves a realistic adversary that carefully designs and inserts malicious chats into the system to break multiple layers of privacy protections and induce the leakage of sensitive information from a target user's chat. We evaluated CLIOPATRA on synthetically generated medical target chats, demonstrating that an adversary who knows only the basic demographics of a target user and a single symptom can successfully extract the user's medical history in 39% of cases by just inspecting Clio's output. Furthermore, CLIOPATRA can reach close to 100% when Clio is configured with other state-of-the-art models and the adversary's knowledge of the target user is increased. We also show that existing ad hoc mitigations, such as LLM-based privacy auditing, are unreliable and fail to detect major leaks. Our findings indicate that even when layered, current heuristic protections are insufficient to adequately protect user data in LLM-based analysis systems.

Metadata

arXiv ID: 2603.09781
Provider: ARXIV
Primary Category: cs.CR
Published: 2026-03-10
Fetched: 2026-03-11 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.09781v1</id>\n    <title>CLIOPATRA: Extracting Private Information from LLM Insights</title>\n    <updated>2026-03-10T15:17:14Z</updated>\n    <link href='https://arxiv.org/abs/2603.09781v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.09781v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>As AI assistants become widely used, privacy-aware platforms like Anthropic's Clio have been introduced to generate insights from real-world AI use. Clio's privacy protections rely on layering multiple heuristic techniques together, including PII redaction, clustering, filtering, and LLM-based privacy auditing. In this paper, we put these claims to the test by presenting CLIOPATRA, the first privacy attack against \"privacy-preserving\" LLM insight systems. The attack involves a realistic adversary that carefully designs and inserts malicious chats into the system to break multiple layers of privacy protections and induce the leakage of sensitive information from a target user's chat.\n  We evaluated CLIOPATRA on synthetically generated medical target chats, demonstrating that an adversary who knows only the basic demographics of a target user and a single symptom can successfully extract the user's medical history in 39% of cases by just inspecting Clio's output. Furthermore, CLIOPATRA can reach close to 100% when Clio is configured with other state-of-the-art models and the adversary's knowledge of the target user is increased. We also show that existing ad hoc mitigations, such as LLM-based privacy auditing, are unreliable and fail to detect major leaks. Our findings indicate that even when layered, current heuristic protections are insufficient to adequately protect user data in LLM-based analysis systems.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <published>2026-03-10T15:17:14Z</published>\n    <arxiv:primary_category term='cs.CR'/>\n    <author>\n      <name>Meenatchi Sundaram Muthu Selva Annamalai</name>\n    </author>\n    <author>\n      <name>Emiliano De Cristofaro</name>\n    </author>\n    <author>\n      <name>Peter Kairouz</name>\n    </author>\n  </entry>"
}