AI safety pioneer Eliezer Yudkowsky to keynote 22nd SBS D Forum

AI safety pioneer Eliezer Yudkowsky will make his first visit to South Korea on Thursday, the 5th, to warn of superintelligence risks at the 22nd SBS D Forum. Held at the Dongdaemun Design Plaza, the event will feature experts discussing human autonomy and the results of an AI public perception survey of 2,000 citizens.

The 22nd SBS D Forum (SDF) will be held on Thursday, the 5th, at the Dongdaemun Design Plaza (DDP) under the theme "In the Age of Substitute Intelligence, Questioning Cognitive Sovereignty." As the era of "agentic AI"—which moves beyond generative AI to independently judge and act—expands, the forum will gather domestic and international experts in AI safety, philosophy, neuroscience, and the arts to examine how much judgment humans should delegate to AI and what is required to remain masters of our own choices and decisions.

The keynote speaker for this forum is Eliezer Yudkowsky, a pioneer in the field of AI safety and alignment who has consistently warned of the dangers of "superintelligence." Making his first visit to South Korea, Yudkowsky will warn of the risks humanity faces on the path toward superintelligence and suggest the role South Korea can play. His message is expected to draw significant interest as he contemplates whether humans can control AI once it surpasses human intelligence and how the AI development race might threaten human survival.

Professor Philip Corlett, founding director of the University of Oxford's Leverhulme Centre for the Future of Intelligence, will address whether humans can maintain autonomy in judgment during the AI era. He will pose the question of whether choices can truly be considered one's own if humans begin to delegate important life decisions to AI. Rather than AI replacing human judgment, Corlett proposes "autonomy-preserving intelligence augmentation," where AI helps humans think better. He is expected to present a direction for how humans can utilize AI to remain the masters of their own judgments.

World-renowned media artist Refik Anadol will provide a visual demonstration of how AI can expand the realm of human creativity. Having used data as paint and AI as a "thinking brush," Anadol has created new forms of art using AI trained on vast amounts of natural data. He will share his thoughts on the possibility of AI evolving beyond a tool for artists to a partner that opens new creative territories with humans.

A session will also address the future of defense, a field where the most acute ethical issues may arise if AI replaces human judgment. Lee Jun-seok, head of the Defense AI Research Institute at Hanwha Aerospace, will present the potential for AI to contribute to national security and AI sovereignty. Im Seong-sin, head of the AI Development Department at Korea Aerospace Industries (KAI), is expected to illustrate a future aerial combat scenario where humans and AI collaborate based on their respective strengths. The session will also examine how much judgment directly linked to human life can be entrusted to AI in an era where AI becomes a core operational tool on the battlefield.

Cha Ji-uk, a professor of psychology at Seoul National University, will introduce the possibility of "Large Brain Models" that connect the human brain with AI. The technology to analyze brain activity to interpret what sentences a person hears and what thoughts they have has the potential to fundamentally change how AI understands humans. The session will look into how close we are to an era where AI understands humans "from the inside out" through the combination of brain models and language models, and the changes this technology will bring.

In the field of culture and arts, bestselling poet Park Jun and Kang Yi-yeon, a chair professor in the Department of Industrial Design at KAIST, will take the stage. Poet Park Jun will discuss the meaning of human memory and places in an AI era that pursues efficiency and speed, touching upon how memories are recalled in present life. Professor Kang Yi-yeon will examine the meaning of human creation and art in the AI era, discussing the value of humanity and subjectivity that technology cannot replace.

The forum will also reveal the results of a large-scale "AI Public Perception Survey" conducted on 2,000 South Korean citizens. This joint research by the Seoul National University Institute for the Future of Strategy, Yonsei University Graduate School of Information, KAIST, and the SBS Foundation diagnoses how AI is changing the daily lives, judgments, and social relationships of Koreans. Discussions will continue regarding how the public sphere will change in the AI era and the role of the media.

The 22nd SBS D Forum will take place on Thursday, the 5th, at the Dongdaemun Design Plaza (DDP) and will be broadcast live on SBS-TV starting at 8:15 AM. The SBS News YouTube channel will provide real-time streaming of all SDF sessions following the TV broadcast until the forum concludes at 5:00 PM.

Copyright © SBS & SBSi. All rights reserved. Unauthorized reproduction, redistribution, or use for AI training is prohibited.

Read the original (Korean)