arXiv:2601.08415cs.CYcs.AI2026-01被引 1

对比五大模型平台条款,揭示研究使用中的法律模糊地带。

Regulatory gray areas of LLM Terms

  • 对比Anthropic等五家厂商条款,发现限制差异大
  • 识别出安全、社科及心理研究中的使用困境
  • 提供公开资源帮助研究者规避合规风险

大型语言模型(LLMs)正日益融入学术研究流程,但其使用条款却长期缺乏深入审视。本文基于2025年11月收集的Anthropic、DeepSeek、Google、OpenAI和xAI五家主流提供商的Terms of Service,开展对比分析。结果显示,普通用户与研究者的使用限制在严格程度和具体性上存在显著差异。研究特别指出,在安全研究、计算社会科学及心理学研究领域存在明确复杂性。我们识别出若干‘监管灰色地带’,即条款导致合法用途产生不确定性。为此,我们贡献了一个公开可访问的比较资源(OSF),并讨论了对一般用户与研究者应对这一动态环境的启示。

原文摘要 · Abstract (English)

Large Language Models (LLMs) are increasingly integrated into academic research pipelines; however, the Terms of Service governing their use remain under-examined. We present a comparative analysis of the Terms of Service of five major LLM providers (Anthropic, DeepSeek, Google, OpenAI, and xAI) collected in November 2025. Our analysis reveals substantial variation in the stringency and specificity of usage restrictions for general users and researchers. We identify specific complexities for researchers in security research, computational social sciences, and psychological studies. We identify `regulatory gray areas' where Terms of Service create uncertainty for legitimate use. We contribute a publicly available resource comparing terms across platforms (OSF) and discuss implications for general users and researchers navigating this evolving landscape.

LLM条款研究合规灰色地带

Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。