Sunghee Yun (He/Him)

Co-founder & CTO @ Erudio Bio, Inc., Chief Business Development Officer @ WeStory.ai, Advisor & Evangelist @ CryptoLab, Inc., KFAS-Salzburg Global Leadership Initiative Fellow @ Salzburg Global Seminar, Adjunct Professor @ Sogang University, Advisory Professor @ DGIST, Global Advisory Board Member @ Innovative Future Brain-Inspired Intelligence System Semiconductor of Sogang University, Network Expert @ Gerson Lehrman Group

Abstract

This presentation explores the transformative landscape of artificial intelligence, examining its rapid evolution from the deep learning revolution of 2012-2015 through the current era of multimodal AI agents. The talk traces AI’s significant milestones, including AlexNet’s breakthrough in computer vision, AlphaGo’s strategic mastery, the transformer architecture’s revolutionary impact on natural language processing, and recent advances in specialized applications like AlphaFold for protein folding and humanoid robotics. Through mathematical analysis of sequence inference probabilities, the presentation explains why we observe sudden dramatic improvements in large language model performance as token accuracy approaches saturation, providing theoretical insight into the current AI capabilities explosion.

The discussion addresses critical questions surrounding AI’s societal implications, including the distinction between human-level performance benchmarks and AI’s unique advantages, the philosophical nature of knowledge, belief, and reasoning in large language models, and the pervasive issue of cognitive biases in both human and machine intelligence. Drawing from cognitive scientific perspectives, the presentation argues that while LLMs excel at conditional probability estimation and can mimic reasoning through chain-of-thought prompting, they fundamentally lack the grounding in shared human experience necessary for true knowledge and belief. This analysis warns against anthropomorphizing AI systems while acknowledging their remarkable utility and commercial potential.

The presentation culminates with insights from the KFAS-Salzburg Global Leadership Initiative, proposing a framework for “reclaiming technology for humanity” through strategic AI development that enhances rather than replaces human capabilities. Key recommendations include comprehensive AI capacity building for scientists, engineers, policymakers, and lawmakers, emphasizing ethics education, bias detection, transparency, and environmental sustainability. The talk advocates for participatory social agreements encompassing data sovereignty, corporate responsibility, and algorithmic impact assessments, ultimately calling for a paradigm shift that channels AI’s transformative power toward human-centric solutions while addressing fundamental challenges of inequality, misinformation, and technological displacement.