arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.05056cs.CRcs.CLcs.LG
收藏

grounded but Misleading: Evaluating Semantic Alignment in AI-Generated Security Explanations

Grounded but Misleading: Evaluating Semantic Alignment in AI-Generated Security Explanations

Virginia Tech(弗吉尼亚理工学院)

Heajun An, Connor Ng, Sandesh Sharma Dulal, Junghwan Kim, Jin-Hee Cho

AI总结:

本文研究了AI生成的安全解释中语义对齐的问题,通过VEXA测试平台验证了词汇基础与语义风险对齐之间的差距,发现即使解释在词汇上显得合理,其语义解释可能削弱检测器的意图风险评估。

摘要:

在线诈骗越来越多地利用流畅且具有上下文意识的社会工程策略,导致对能够解释为何一条信息可能具有风险的AI系统的需求日益增长。然而,引用检测器衍生证据的解释可能仍然在语义上削弱或改变预期的风险解释。我们介绍了VEXA:验证语义解释对齐,一个用于研究AI生成诈骗风险解释中词汇基础与语义风险对齐差距的受控测试平台。VEXA通过独立控制证据基础和语义框架来生成无基础、风险对齐和风险稀释的解释。通过LLM作为判断者和人类评估,我们发现即使解释的语义解释削弱了检测器的意图风险评估,解释仍可能在比较上显得合理。在人类评估中,风险稀释的XAI基础解释保留了相对较高的感知证据基础评分(3.66),尽管其帮助性(3.00)和推理支持(3.14)评分较低。这些发现提供了AI生成安全解释中基础错觉效应的受控证据,并表明可信的解释评估必须不仅验证是否引用了证据,还要验证如何解释这些证据。

英文摘要:

Online scams increasingly leverage fluent and context-aware social engineering strategies, creating growing demand for AI systems that explain why a message may be risky. However, explanations that cite detector-derived evidence may still semantically weaken or redirect the intended risk interpretation. We introduce VEXA: Verifying Semantic Explanation Alignment, a controlled testbed for studying the gap between lexical grounding and semantic risk alignment in AI-generated scam-risk explanations. VEXA generates ungrounded, risk-aligned, and risk-diluting explanations by independently controlling evidence grounding and semantic framing. Through LLM-as-a-judge and human evaluations, we show that explanations may continue to appear comparatively grounded even when their semantic interpretation weakens the detector's intended risk assessment. In human evaluation, risk-diluting XAI-grounded explanations retained comparatively elevated Perceived Evidence Grounding scores (3.66) despite lower Helpfulness (3.00) and Reasoning Support (3.14) scores. These findings provide controlled evidence of grounding illusion effects in AI-generated security explanations and suggest that trustworthy explanation evaluation must verify not only whether evidence is cited, but also how that evidence is interpreted.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑