Red Teaming for Generative AI, Report on a Copyright-Focused Exercise Completed in an Academic Medical Center

James Wen; Sahil Nalawade; Zhiwei Liang; Catherine Bielick; Marisa Ferrara Boston; Alexander Chowdhury; Adele Collin; Luigi De Angelis; Jacob Ellen; Heather Frase; Rodrigo R. Gameiro; Juan Manuel Gutierrez; Pooja Kadam; Murat Keceli; Srikanth Krishnamurthy; Anne Kwok; Yanan Lance Lu; Heather Mattie; Liam G. McCoy; Katherine Miller; Allison C. Morgan; Marlene Louisa Moerig; Trang Nguyen; Alexander Owen-Post; Alex D. Ruiz; Sreekar Reddy Puchala; Soujanya Samineni; Takeshi Tohyama; Varun Ullanat; Carmine Valenza; Camilo Velez; Pengcheng Wang; Anna Wuest; Yuxiang Zhou; Yingde Zhu; Jason M. Johnson; Naomi Lenane; Jennifer Willcox; Francis J. Vitiello; Leo Anthony G. Celi; Renato Umeton

arXiv· arXiv:2506.22523· James Wen, Sahil Nalawade, Zhiwei Liang, Catherine Bielick, Marisa Ferrara Boston, Alexander Chowdhury, Adele Collin, Luigi De Angelis, Jacob Ellen, Heather Frase, Rodrigo R. Gameiro, Juan Manuel Gutierrez, Pooja Kadam, Murat Keceli, Srikanth Krishnamurthy, Anne Kwok, Yanan Lance Lu, Heather Mattie, Liam G. McCoy, Katherine Miller, Allison C. Morgan, Marlene Louisa Moerig, Trang Nguyen, Alexander Owen-Post, Alex D. Ruiz, Sreekar Reddy Puchala, Soujanya Samineni, Takeshi Tohyama, Varun Ullanat, Carmine Valenza, Camilo Velez, Pengcheng Wang, Anna Wuest, Yuxiang Zhou, Yingde Zhu, Jason M. Johnson, Naomi Lenane, Jennifer Willcox, Francis J. Vitiello, Leo Anthony G. Celi, Renato Umeton·本站收录2026-10-29 06:17· 原文发表2025年6月27日

生成式 AI 红队测试：Dana-Farber 癌症研究所的版权专项测试报告

Red Teaming for Generative AI, Report on a Copyright-Focused Exercise Completed in an Academic Medical Center

AI 导读

Dana-Farber 癌症研究所对其基于 OpenAI 模型的内部生成式 AI 工具 GPT4DFCI 开展了版权合规红队测试。2024 年 11 月，42 名来自学术、产业界和政府机构的参与者组成四个团队，尝试从文学作品、新闻文章、科学论文和访问受限的临床笔记四类内容中提取受版权保护材料。测试成功逐字提取书籍献词并获得近乎原文的段落；新闻文章在越狱尝试下仍未能提取，科学论文仅复现出高层摘要，临床笔记测试显示隐私防护有效。作者指出，文学作品提取成功提示训练数据中可能存在受版权材料，需要推理时过滤，不同内容类型的成功率差异反映防护机制不同。此次测试促成 GPT4DFCI 上线版权专用元提示词，自 2025 年 1 月起投入生产。

阅读原文arxiv.org