Flitto CEO: AI is breaking down language barriers

By AJP Posted : December 4, 2024, 17:29 Updated : December 4, 2024, 17:29
Lee Jung-soo, CEO of Flitto, speaks during the launch ceremony of AJP in Seoul's FKI Tower on Dec. 3, 2024. AJP Yoo Dae-gil
SEOUL, December 4 (AJP) - "By 2025, the language barriers people experience today will be significantly reduced," says Lee Jung-soo, CEO of Flitto, predicting that an influx of new services will transform various aspects of society and economy.

Speaking at the launch of the Asian news platform AJP in Seoul's FKI Tower on Tuesday, Lee outlined how artificial intelligence is dismantling language barriers that humanity has long struggled to overcome.

Lee envisions a future where AI technology will make previously inaccessible information available to everyone, transforming everything from job hunting to international shopping. 

"People will be able to shop freely on foreign language e-commerce sites, and language restrictions in recruitment processes will disappear," he said.

This breakthrough, Lee explained, has been made possible by advances in communication speeds, GPU technology, and the emergence of Large Language Models (LLM) trained on vast linguistic datasets.

"The Universal Translator from Star Trek series in the 2000s resembled today's smartphones. While rapid elimination of language barriers was predicted then, only recent technological advances have made it feasible," he said.

However, Lee pointed out several remaining challenges. 

"One major issue is understanding language forms. There are over 7,000 languages worldwide," he said, noting that current AI systems struggle with context-dependent language variations, such as gender-specific terms and honorific expressions.

Proper noun recognition presents another significant challenge. "When you ask 'Where is Dongdaemun?' in English on Google Translate, it sometimes fails to recognize 'Dongdaemun' as a proper noun. While countless proper nouns exist, accurate AI recognition remains difficult," Lee explained.

The solution, according to Lee, lies in continuous advancement of AI data technology. "Just as humans need constant learning, AI requires continuous data input and refinement. Mastery of AI technology essentially means excelling at efficient data management," he said.

Lee concluded by emphasizing the importance of human-AI collaboration, stating that human refinement and retraining are crucial for addressing AI's current limitations.

The presentation was part of the launch event for AJP, a new English-language news platform focused on Asian perspectives.

Copyright ⓒ Aju Press All rights reserved.