在线有害内容应对策略分类：证据评估 (A Taxonomy of Response Strategies to Toxic Online Content: Evaluating the Evidence)

Toxic Online Content (TOC) includes messages on digital platforms that are harmful, hostile, or damaging to constructive public discourse. Individuals, organizations, and LLMs respond to TOC through counterspeech or counternarrative initiatives. There is a wide variation in their goals, terminology, response strategies, and methods of evaluating impact. This paper identifies a taxonomy of online response strategies, which we call Online Discourse Engagement (ODE), to include any type of online speech to build healthier online public discourse. The literature on ODE makes contradictory assumptions about ODE goals and rarely distinguishes between them or rigorously evaluates their effectiveness. This paper categorizes 25 distinct ODE strategies, from humor and distraction to empathy, solidarity, and fact-based rebuttals, and groups these into a taxonomy of five response categories: defusing and distracting, engaging the speaker's perspective, identifying shared values, upstanding for victims, and information and fact-building. The paper then systematically reviews the evidence base for each of these categories. By clarifying definitions, cataloging response strategies, and providing a meta-analysis of research papers on these strategies, this article aims to bring coherence to the study of ODE and to strengthen evidence-informed approaches for fostering constructive ODE.

翻译：在线有害内容包括数字平台上对建设性公共讨论具有危害性、敌对性或破坏性的信息。个人、组织及大语言模型通过反驳性言论或对抗性叙事举措应对此类内容。这些应对措施在目标设定、术语使用、响应策略及效果评估方法上存在显著差异。本文提出一种在线响应策略分类体系——在线话语参与，涵盖所有旨在构建更健康网络公共讨论的在线言论形式。现有文献对在线话语参与的目标存在矛盾假设，鲜有研究对其目标进行明确区分或严格评估其有效性。本文系统归类了25种不同的在线话语参与策略（从幽默转移注意力到共情支持、基于事实的反驳等），并将其归纳为五大应对类别：缓和与转移注意力、介入发言者视角、寻求共同价值观、维护受害者立场、信息与事实构建。随后，本文对各策略类别的证据基础进行了系统性综述。通过厘清定义、梳理应对策略、对相关研究文献进行元分析，本文旨在增强在线话语参与研究的体系化，并为促进建设性在线话语参与提供循证实践路径。