▲ A scene of the AI interpretation program being used at an international symposium.

International academic conferences in specialized fields such as medicine often involve complex terminology and abbreviations, making interpretation particularly challenging. Conventional AI-based translation tools frequently fall short in accuracy due to insufficient contextual understanding. Moreover, translation usually begins only after the speaker finishes speaking, making it difficult for attendees to follow the content in real time.

Professors Yong-Gil Kim and Ji Seon Oh of the Division of Rheumatology at Asan Medical Center, together with Professor Woo Hyun Shim of the Biomedical Research Center at Asan Medical Center, have recently developed an AI interpretation program designed for real-time use at international conferences across various specialized fields. The program, trained on domain-specific terminology in areas such as medicine and law, analyzes spoken data and provides natural, contextually appropriate translations. It translates speech word by word in real time, displaying complete sentences on screen within one to three seconds, allowing attendees to quickly understand the content in their native language.

The program was first introduced at the Korean College of Rheumatology International Symposium, held for three days from May 15, received positive feedback from over 900 attendees from 27 countries. The research team has recently filed technology patents related to this development, including a multilingual translation system, a device for processing real-time speech recognition and translation data, and a domain-specific translation systems and techniques utilizing multimodal information.