analiza aiqbench
This paper introduces RubRIX, a novel evaluation framework specifically designed to assess nuanced safety risks in large language models used for caregiving contexts. By moving beyond generic toxicity checks, it quantifies domain-specific dangers like bias and epistemic arrogance across six state-of-the-art models using over 20,000 real-world queries. The study demonstrates that rubric-guided refinement can significantly reduce these interactional risks, offering a critical methodological advance for deploying AI in high-stakes social support applications.
RubRIX: Rubric-Driven Risk Mitigation in Caregiver-AI Interactions
arXiv:2601.13235v2 Announce Type: replace-cross Abstract: 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.
Abstract: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.
Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Machine Learning (cs.LG); Social and Information Networks (cs.SI)
arXiv:2601.13235 [cs.HC]
(or arXiv:2601.13235v2 [cs.HC] for this version)
https://doi.org/10.18653/v1/2026.findings-acl.1774 DOI(s) linking to related resources
Text integral sub licența deschisă a publicației · 200 cuvinte