The societal impact of large AI systems is unprecedented. Researchers across various fields actively contribute to examining and addressing potential AI harm by employing novel methods that not only uncover possible risks but also develop solutions to mitigate them before deploying an AI system. These methods range from red-teaming AI to gathering feedback from multi-stakeholder groups, as well as experimenting with diverse approaches to model training and evaluation. In this talk, Angel Hsing-Chi Hwang will provide an overview of these methods and explore whether these large AI systems per se can serve as tools for researchers and practitioners to assess AI’s societal impact. Specifically, Hwang will discuss case studies that apply LLMs in participatory processes and simulation-based research. Hwang would also like to engage in discussions and hear from fellow researchers in the audience about their views on the potential and concerns regarding the use of these approaches to address various societal challenges related to their research.

Hwang (she/her) is an assistant professor at USC Annenberg and a researcher in human-AI interaction (HAII). Her research explores the societal impact of artificial intelligence (AI) on work practices across various fields, including small-group teamwork, content production, and mental healthcare services. Additionally, her work aims to provide insights for practitioners to design, develop, and implement AI-powered technologies that promote better work outcomes. Hwang utilizes a diverse range of methods to investigate these topics, including behavioral experiments, data-driven research, and qualitative approaches. She is also interested in exploring novel (combinations of) methods that can more effectively predict the impact of AI applications before large-scale deployment.

This program is open to all eligible individuals. USC Annenberg operates all of its programs and activities consistent with the University’s Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor.