Guangwei Zhang, Jianing Zhu, Cheng Qian, Neil Gong, Rada Mihalcea, Zhaozhuo Xu, Jingrui He, Jiaqi Ma, Yun Huang, Chaowei Xiao, Bo Li, Ahmed Abbasi, Dongwon Lee, Heng Ji, Denghui Zhang
Copyright Detective is an interactive forensic system designed to detect and analyze potential copyright risks in the outputs of large language models (LLMs).
Copyright Detective is a tool that helps identify and understand potential copyright issues in the text generated by large language models, like those used in AI chatbots. Instead of just labeling text as problematic or not, it treats the process like a detective work, gathering evidence to see if copyrighted material is being improperly used. The system uses several techniques to check if the AI is recalling content too closely or paraphrasing it in a way that still violates copyright. This helps ensure that these AI systems are used responsibly and transparently, even when we can't see inside them.
