arXiv cs.CY· Eli Min, Mohit Chandra, Aamogh Sawant, Mengting Tu, Freya Lin, Ridwan Shams Khan, Reigna El-Yashruti, Christopher L. Smith, Mark Kalinich, John Torous, Munmun De Choudhury·· 11 小时前AI 评分44
记忆还不够：评估 AI 心理健康支持的长期时机与遗漏问题
Remembering Is Not Enough: Evaluating Timing and Omission in Longitudinal AI Mental Health Support
AI 导读
研究基于 22 位用户捐赠的聊天记录构建 14,256 次纵向模拟，考察带记忆的 AI 助手在自杀意念、自残与焦虑危机中的长期支持表现。团队与心理健康专业人员共同提出 CARE 框架评估建议时机与遗漏，发现危机后建议与用户准备度匹配更差，主要源于过早建议，且在恢复停滞时最常见。显性无效建议虽少，但助手常遗漏随访与下一步规划等临床预期支持。
正文 · 原文
Authors:Eli Min, Mohit Chandra, Aamogh Sawant, Mengting Tu, Freya Lin, Ridwan Shams Khan, Reigna El-Yashruti, Christopher L. Smith, Mark Kalinich, John Torous, Munmun De Choudhury
Abstract:Evaluations of large language models (LLMs) for mental health support commonly focus on the clinical appropriateness of individual responses, potentially overlooking failures that emerge across extended interactions. As AI assistants accumulate memory of users' circumstances and prior disclosures, this raises an accountability question: Do AI chatbots provide appropriate support as users' needs evolve? We investigate by analyzing 14,256 longitudinal simulations grounded in donated chatbot histories from 22 users, examining how a memory-equipped assistant responds to suicidal ideation, self-harm, and anxiety crises across recovery trajectories and communication styles. We introduce the Context-Aware and Readiness-Evaluative (CARE) framework, developed with mental health professionals, to assess the timing and quality of advice, including omissions. Advice was less aligned with users' readiness after a crisis than in matched control stretches of the same conversation, largely because of premature recommendations, which were most common when recovery stalled. Revisiting a crisis reduced premature advice but not lagging advice, and users' communication style changed how often and when the assistant gave recommendations. Although explicitly unproductive advice was rare, the assistant frequently omitted clinician-expected support such as follow-up and planning for next steps. These findings show the limits of single response-level safety evaluation and the need to assess timing, omission, and consistency across users when holding AI systems accountable for longitudinal support.
| Subjects: | Computers and Society (cs.CY); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.08721 [cs.CY] |
| (or arXiv:2611.08721v1 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08721 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Elizabeth Min [view email]
[v1]
Thu, 5 Nov 2026 21:14:16 UTC (454 KB)
来源：arXiv cs.CY · arxiv.org