"Building Trustworthy AI Through Collaboration" - Qing Zeng

Abstract: Artificial intelligence is increasingly used in healthcare, but building tools that people trust requires close collaboration across many groups. This talk describes several projects that illustrate how collaborative approaches can support the development of trustworthy AI in practice. Two AIM-AHEAD–funded projects, AI-FOR-U and ArtAI, focus on community-centered collaboration. AI-FOR-U, developed with a historically Black college and university (HBCU), aims to build reliable AI tools for resource-limited settings while developing local expertise throughout the process. ArtAI is a collaboration with SPARC, a nonprofit serving individuals with severe and multiple disabilities, and explores how generative AI can support therapeutic art activities through co-design with participants and caregivers. The talk also discusses Medication-Wide Association Study Plus (MWAS+) as an example of large-scale analytic work where trust depends on transparent methods and collaboration between data scientists, clinicians, and other domain experts. Together, these examples show how collaboration can help produce AI systems that are understandable, credible, and useful across different healthcare contexts. Bio: Dr. Qing Zeng is a tenured professor in the Department of Clinical Research and Leadership at The George Washington University and a research scientist at the Washington, DC Veterans Affairs (VA) Medical Center. She is the Director of the Biomedical Informatics Center at GWU and Co-Director of the Center for Data Science Outcomes Research at DCVA and co-leads the informatics core of the GW–CN CTSI. With decades of experience in biomedical informatics, Dr. Zeng’s work spans data mining, consumer health informatics, and semantic integration across heterogeneous data sources. She has authored over 250 peer-reviewed publications and has served as PI or co-PI on numerous projects funded by VA HSR&D, NIH, CDC, and the Department of Defense.