Empowering Responsible Use of Large Language Models

Explore the responsible use of large language models in this informative lecture by Lei Li from Carnegie Mellon University. Delve into two critical questions surrounding LLMs: identifying AI-generated text and safeguarding intellectual property. Learn about innovative watermarking techniques for both text and models, including their theoretical foundations and robustness against evasion attacks. Gain insights into empirical evidence supporting their effectiveness, providing a solid technical foundation for policymakers, legal professionals, and AI practitioners. Discover the speaker's background as an Assistant Professor at CMU's Language Technology Institute, his notable achievements including the ACL 2021 Best Paper Award, and his experience founding ByteDance AI Lab and developing widely-used AI systems.