Skip main navigation

Bridging Global and Korean AI

Global models do not know Korean well because most training data is English. Watch Professor Hosung Jo explain more.

This chapter analyzes the linguistic and cultural limitations of global AI models trained predominantly on English data. It focuses on fine-tuning and lexical adaptation techniques that enhance the models’ ability to handle Korean grammar, honorifics, and contextual nuance. The chapter further examines hybrid architectures that integrate global reasoning with Korean-specific linguistic knowledge to achieve more accurate, natural, and culturally coherent language processing.

What you will learn in this video:

  • Limitations of global AI models for Korean language processing
  • Fine-tuning and lexical adaptation techniques
  • Hybrid AI systems combining global and local models
  • Future development of Korean-adaptive AI
This article is from the free online

AI and Tech in Korea: A Cultural and Technological Perspective

Created by
FutureLearn - Learning For Life