MRMMIA: Membership Inference Attacks on Memory in Chat Agents
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911723445813248 |
|---|---|
| author | Chen, Kai Pang, Yan Wang, Tianhao |
| author_facet | Chen, Kai Pang, Yan Wang, Tianhao |
| contents | Membership inference attacks (MIAs) test whether a target data record belongs to a system's private data, and have become a standard tool to measure privacy leakage in machine learning systems. Prior work has primarily focused on training corpora or retrieval databases. However, MIAs against agent memory have received less attention, even though such memory can contain sensitive user-agent interactions, retrieved facts, and user preferences. Therefore, in this work, we focus on chat agent memory MIAs, where an adversary infers whether a candidate memory unit belongs to the chat agent's memory store. We propose Multi-Recall Memory MIA (MRMMIA), a unified attack that utilizes multiple recall probes to the agent to extract the membership signal across black-box, gray-box, and white-box settings. Our experiments demonstrate that MRMMIA consistently outperforms baselines. Our results expose the privacy risk in agents and provide an initial evaluation framework for membership leakage in chat-agent memory systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_27825 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MRMMIA: Membership Inference Attacks on Memory in Chat Agents Chen, Kai Pang, Yan Wang, Tianhao Cryptography and Security Machine Learning Membership inference attacks (MIAs) test whether a target data record belongs to a system's private data, and have become a standard tool to measure privacy leakage in machine learning systems. Prior work has primarily focused on training corpora or retrieval databases. However, MIAs against agent memory have received less attention, even though such memory can contain sensitive user-agent interactions, retrieved facts, and user preferences. Therefore, in this work, we focus on chat agent memory MIAs, where an adversary infers whether a candidate memory unit belongs to the chat agent's memory store. We propose Multi-Recall Memory MIA (MRMMIA), a unified attack that utilizes multiple recall probes to the agent to extract the membership signal across black-box, gray-box, and white-box settings. Our experiments demonstrate that MRMMIA consistently outperforms baselines. Our results expose the privacy risk in agents and provide an initial evaluation framework for membership leakage in chat-agent memory systems. |
| title | MRMMIA: Membership Inference Attacks on Memory in Chat Agents |
| topic | Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2605.27825 |