Chaowei Xiao, an assistant professor in the Information School, has been named an AI2050 Early Career Fellow by Schmidt Sciences, a philanthropy dedicated to fostering the advancement of science and technology. The selection cements Xiao’s place among the world’s top young scholars in AI safety and security.
According to Schmidt Sciences (co-founded by former Google CEO Eric Schmidt), the AI2050 initiative envisions a world where in 2050, AI has been “hugely beneficial to humanity.” In this scenario, the researchers are setting out to answer the retrospective question, “What are the most important problems we solved and the opportunities and possibilities we realized to ensure this outcome?”
Through the prestigious two-year fellowship, offered to just 20 outstanding early-career researchers from around the world, Xiao will advance his research and expertise at the intersection of machine learning and security, with the goal of building secure and safe machine learning systems for the future. “Being named a Schmidt Early Career Fellow is an extremely great honor, and I’m excited to join such an impressive cohort of AI2050 fellows,” Xiao said. “It will enable me to work with my students to build next-generation safe and secure AI systems.”
More specifically, Xiao’s project aims to create advanced tools that can automatically find and analyze weaknesses in AI systems, through a process known as red-teaming. “Think of it like an antivirus tool, but for AI systems,” he said. Additionally, Xiao and his collaborators will identify and assess other emerging harms caused by AI, in an effort to ensure future systems become safer, more robust, and broadly beneficial to society.
Xiao was inspired to pursue AI safety research when during his PhD research, he witnessed an AI model for self-driving vehicles misreading stop signs, a flaw that could lead directly to physical harm if deployed in the real world. This moment spurred him to research these systems and dedicate his career to “making AI better for our collective lives,” he said.
Congratulations to Chaowei Xiao on this distinguished honor!
For more on Chaowei Xiao’s AI2050 project, visit this page.
To learn more about Chaowei Xiao’s research, view his website.