UC San Diego Joint AI/Safety/Health Events -- February 19-20 2025

The rapid advancements in artificial intelligence (AI) have transformed numerous industries, but they also bring a host of challenges that demand urgent attention. The proposed two-day event at UCSD is co-organized by a number of research centers, institutes, and departments at UCSD along with their partners. The events are scheduled for February 19-20, 2025, aims to address some of the most critical issues facing AI today, with a particular focus on safety, robustness, security, and privacy. The event will bring together researchers, industry leaders, and policymakers to engage in thoughtful discussions and collaborative problem-solving. By fostering interdisciplinary dialogue, this event seeks to identify actionable pathways for making AI systems safer and more reliable, especially in high-stakes domains such as healthcare and biomedical applications.

The event is structured to address key thematic areas, each with its own dedicated day. The first day will focus on AI safety, exploring the foundational principles and techniques required to ensure AI systems behave as intended, even in unpredictable or adversarial environments. The second day will delve deeper into the safety, robustness, security, and privacy of AI in health and medical applications, a field where the stakes are particularly high due to the sensitivity of patient data and the potential for life-altering decisions. We will have speakers from the defense side, drawing parallels between secure AI in defense and health domains, as well as other critical application domains. Finally, the third day will broaden the conversation to examine the challenges of integrating AI into bio and health applications, tackling issues like interoperability, regulatory compliance, and the ethical implications of AI in these transformative fields. Together, these sessions will provide a comprehensive framework for addressing the pressing issues at the intersection of AI and healthcare, fostering innovations that are both impactful and trustworthy.