Jing Shao
Jing Shao is a Research Scientist with Shanghai Artificial Intelligence Laboratory, China. She is Co-Director of the Center of Safe & Trustworthy AI. She is also an Adjunct Ph.D. supervisor at Shanghai Jiao Tong University and Fudan University. She received her Ph.D. (2016) from Multimedia Lab (MMLab), The Chinese University of Hong Kong (CUHK), supervised by Prof. Xiaogang Wang, and worked closely with Prof. Chen Change Loy and Prof. Xiaoou Tang. Prior to joining Shanghai AI Lab, she served as a Research Director at SenseTime from 2016 to 2023.
Her research interests focus on multi-modal foundation models and agents, with special interests in understanding various properties of current models beyond their accuracy, such as safety, explainability, robustness, and generalization, towards the reliableness of AI systems. She has published 50+ peer-reviewed articles in top-tier conferences and journals such as TPAMI, IJCV, ICML, ICLR, NeurIPS, CVPR, and ACL, with 11500+ citations in Google Scholar. She is recognized as Stanford Top 2% Worldwide Scientist in 2024, and received ACL 2024 Outstanding Paper Award.
To Prospective Students: We are actively hiring Full-time Researchers/Interns working together on safety/robustness/explainablity of generative models and agents. I am also looking for talented students targeted to Master or Ph.D. degree. Please drop me an email with your resume if you are interested.
News
| Aug 21, 2025 | 5 papers accepted to EMNLP 2025! |
|---|---|
| Jul 31, 2025 | ACL 2025 Outstanding Paper Award: “LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts” |
| Jul 27, 2025 | The first frontier AI risk management framework and comprehensive implementation report applicable to general-purpose AI models and agents: Framework, Practice |
| Jun 30, 2025 | We are looking for talented 2026 Ph.D. Students |
| Aug 14, 2024 | ACL 2024 Outstanding Paper Award: “PsySafe: A Comprehensive Framework for Psychological-based Attack, Defense, and Evaluation of Multi-agent System Safety” |
Selected Recent Publications
2024
-
PsySafe: A Comprehensive Framework for Psychological-based Attack, Defense, and Evaluation of Multi-agent System SafetyACL 2024, 2024
2024
-
LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution ShiftsACL 2025, 2024
2024
-
REEF: Representation Encoding Fingerprints for Large Language ModelsICLR 2025, 2024
