arXiv cs.AI· Hobin Kim, Xiaoyuan Wu, Omer Akgul, Lujo Bauer, Nicolas Christin·· 1 天前AI 评分40
研究：用户向 LLM 提问的数字安全与隐私问题及模型回答质量
Security and Privacy Prompts in the Wild: What Users Ask LLMs and How LLMs Respond
AI 导读
一项基于 WildChat 中 320 万条真实对话的研究识别出 14,727 条数字安全与隐私（S&P）提示词，并归纳为九类。在 270 条求助类提示词上各重复提问 10 次后，GPT 5.5 有 98% 的回答达到"足够好"，Llama 4 仅为 47%。商业模型虽平均质量更高，但在多次运行间有时给出语义不一致甚至相互矛盾的回答。
正文 · 原文
Abstract:Large language models (LLMs) are widely used to answer users' information needs: users ask LLMs about the weather, pose educational questions, and consult them for legal assistance. One particularly understudied area is digital security and privacy (S&P), where users may seek LLMs' help on how to secure their online accounts or protect their computers from cyber attacks. To the best of our knowledge, no prior study has collected or analyzed the S&P questions users ask LLMs; prior research on LLM response quality relied on expert-authored S&P misconceptions or FAQs rather than user queries. Drawing from WildChat, a dataset of 3.2M user-LLM conversations collected in the wild, our study identifies 14,727 S&P prompts and categorizes them into nine categories covering a wide range of S&P topics. We sampled 450 S&P prompts, and performed a thematic analysis to characterize the S&P questions users ask LLMs. Separate from the thematic analysis, we curated 270 advice-seeking S&P prompts, where users ask for recommendations, guidance, or specific S&P information. We measured LLM response quality and consistency when posing each prompt to the LLMs 10 times. We found that commercial LLMs outperform open-weight models (GPT 5.5 provided "good enough" responses on 98% of prompts; Llama 4 on 47%). However, among prompts that received high-quality responses on average, commercial models sometimes produce semantically inconsistent responses across runs (some are contradictory, and others are different in specificity), thereby risking confusing or misleading users.
| Comments: | Published as a main conference paper at EMNLP 2026 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Human-Computer Interaction (cs.HC) |
| Cite as: | arXiv:2606.18062 [cs.CL] |
| (or arXiv:2606.18062v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.18062 arXiv-issued DOI via DataCite |
|
| Related DOI: | https://doi.org/10.18653/2026.emnlp-main.2571
DOI(s) linking to related resources |
Submission history
From: Xiaoyuan Wu [view email]
[v1]
Tue, 16 Jun 2026 15:37:25 UTC (486 KB)
[v2]
Tue, 10 Nov 2026 21:37:52 UTC (502 KB)
来源：arXiv cs.AI · arxiv.org