Caregivers seeking AI-mediated support express complex needs -- information-seeking, emotional validation, and distress cues -- that warrant careful evaluation of response safety and appropriateness. Existing AI evaluation frameworks, primarily focused on general risks (toxicity, hallucinations, policy violations, etc), may not adequately capture the nuanced risks of LLM-responses in caregiving-contexts. We introduce RubRIX (Rubric-based Risk Index), a theory-driven, clinician-validated framework for evaluating risks in LLM caregiving responses. Grounded in the Elements of an Ethic of Care, RubRIX operationalizes five empirically-derived risk dimensions: Inattention, Bias & Stigma, Information Inaccuracy, Uncritical Affirmation, and Epistemic Arrogance. We evaluate six state-of-the-art LLMs on over 20,000 caregiver queries from Reddit and ALZConnected. Rubric-guided refinement consistently reduced risk-components by 45-98% after one iteration across models. This work contributes a methodological approach for developing domain-sensitive, user-centered evaluation frameworks for high-burden contexts. Our findings highlight the importance of domain-sensitive, interactional risk evaluation for the responsible deployment of LLMs in caregiving support contexts. We release benchmark datasets to enable future research on contextual risk evaluation in AI-mediated support.

翻译：寻求人工智能辅助支持的照护者表达出复杂需求——包括信息获取、情感认同与压力信号——这些需求要求对响应的安全性与适宜性进行审慎评估。现有的人工智能评估框架主要关注通用风险（如有害内容、幻觉、政策违规等），可能无法充分捕捉大语言模型在照护情境下响应的细微风险。本文提出RubRIX（基于量表的风险指数），这是一个理论驱动、经临床验证的框架，用于评估大语言模型在照护响应中的风险。该框架以“关怀伦理要素”为理论基础，将五个经验推导的风险维度操作化：关注缺失、偏见与污名化、信息不准确、无批判性认同以及认知傲慢。我们基于来自Reddit和ALZConnected平台的超过20,000条照护者查询，对六个前沿大语言模型进行了评估。经过量表引导的优化后，所有模型在单次迭代中均实现了风险成分持续降低45-98%。本研究为开发面向高负担情境的领域敏感、以用户为中心的评估框架贡献了方法论路径。我们的研究结果凸显了领域敏感、交互式风险评估对于在照护支持场景中负责任地部署大语言模型的重要性。我们公开了基准数据集，以促进未来关于人工智能辅助支持中情境化风险评估的研究。