Events

Value Alignment and Superalignment in AI: Unpacking Safety Risks and Future Directions

Speaker: JinYeong Bak

Location: 1 MetroTech Center, Room 22nd Floor : RSVP at Link in Synopsis

Date: Monday, August 25, 2025

Abstract: As AI systems grow increasingly powerful and integrated into society, aligning them with human values becomes both a technical challenge and a moral imperative. This talk explores the value alignment and superalignment in large language models (LLMs) and future artificial superintelligence (ASI). Drawing from recent empirical and psychological research, we first examine how personalized value alignment in LLMs can amplify harmful behaviors. We then shift focus to the broader horizon of superalignment, arguing that aligning future ASI systems with human values is not only feasible but urgent. By introducing a framework based on alternating optimization of competence and conformity, we propose a path forward for responsible AI development. This talk aims to provoke discussion on the unintended consequences of current alignment strategies and inspire proactive research toward safer, value-aligned AI systems.

Bio: JinYeong Bak is an associate professor in the Colleague of Computing at Sungkyunkwan University. His research interests include analyzing human conversational behaviors and building machine learning models from the insights of the analysis. He worked at Microsoft Research Asia as a research intern and United Nations Global Pulse Lab Jakarta as a junior data scientist. He holds a Ph.D. and M.S. from the KAIST and a B.S. from Sungkyunkwan University. His research has been published in ACL, EMNLP, NAACL, ICML, CHI, and WWW. His personal homepage: https://nosyu.kr and lab homepage: https://hli.skku.edu

Dinner & networking will begin at 6:00 PM and the seminar will start at 7:00 PM EST.. Please RSVP by filling out this Google Form. For online attendees, a Zoom link will be sent out prior to the event.