arXiv cs.CL· Andrea Mock, Valerie Chen, Eric Lin, Rayan Krishnan, Diyi Yang, Caroline Figueroa·· 12 小时前AI 评分48
SAFE-Teen：评估聊天机器人在青少年心理健康对话中的安全性
SAFE-Teen: Evaluating Chatbot Safety Across Adolescent Mental Health Conversations
AI 导读
SAFE-Teen 对 18 个聊天机器人 API 的 32,555 段模拟青少年（13-17 岁）对话进行临床医生参与的评测，基于 22 项安全标准，整体失败率为无 teen-aware 提示词时的 43.3%、加入提示词后的 24.9%。
正文 · 原文
Abstract:Adolescents are increasingly turning to AI chatbots for mental health support, advice, and companionship, raising safety concerns. While recent evaluations have started to examine safety across multiple conversational turns, few compare how slight changes in a teen's responses can affect a chatbot's safety in the same scenario. We conduct a clinician-informed evaluation of chatbot safety in simulated conversations with adolescents aged 13-17, examining how safety failures vary with adolescents' responses and teen-aware system prompting. We evaluated 32,555 simulated conversations across 18 chatbot APIs and two system-prompt conditions using 22 safety criteria grounded in previously identified harms. Conversations spanned 983 variants of 106 seed scenarios covering urgent safety issues, health advice, and identity and relationships. The overall criterion failure rate was 43.3% without the teen-aware prompt and 24.9% with it. The teen-aware prompt reduced failures, but it didn't eliminate them. When a teenager pushed back on a chatbot's answer, failures on mental health diagnosis, medication, and substance misuse boundaries rose by 10-12 percentage points without the prompt, for more than 80% of chatbots. These findings reveal persistent safety failures in simulated adolescent-chatbot interactions and show that safety depends on how a conversation unfolds.
| Comments: | 29 pages. Content warning: this paper discusses sensitive content, including self-harm, suicide, and other mental health topics |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.07218 [cs.CL] |
| (or arXiv:2611.07218v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.07218 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Andrea Mock [view email]
[v1]
Wed, 4 Nov 2026 23:11:33 UTC (110 KB)
来源：arXiv cs.CL · arxiv.org