Puzzle AI Unveils Medical Speech Recognition Technology at InterSpeech 2025

by Kim Jung -ah jungya@chosun.com

08.25.2025 11:35

| Operates even in noisy operating rooms and emergency rooms… Introducing a hands- free medical record solution

AI healthcare startup Puzzle AI (CEO Yong-sik Kim) announced on the 25th that it participated in 'Interspeech 2025,' an international speech technology conference held in Rotterdam, Netherlands, and presented its medical-specialized speech recognition technology. Interspeech is one of the largest international academic events in the field of speech and language processing, where the latest research results are shared every year.

▲ Image provided by Puzzle AI

At this conference, Research Director Hwang Ji-wook and Researcher Kim Do-hyun introduced the Keyword Spotting (KWS) technology applied to Puzzle AI's medical record solution, 'GEN'. This technology recognizes commands such as 'start recording' and 'add test,' enabling medical staff to control electronic medical records without using their hands. In particular, it is characterized by its stable operation even in noisy environments such as operating rooms, emergency rooms, and intensive care units.

According to Puzzle AI, 'GEN' is an all-in-one platform that supports everything from voice input to summarization, generation, and document completion, and is currently being used in over 200 hospitals and clinics, including university hospitals in Korea. The company explained that this presentation at the conference served as an opportunity to internationally showcase the performance of the solution currently being utilized in actual clinical environments.

CEO Kim Yong-sik said, “Hands-free medical records can contribute to reducing the record-keeping burden on medical staff and extending patient consultation time,” adding, “We will strengthen our technological competitiveness in the global medical AI market through this presentation at the conference.”

Source: Digital Chosun(https://digitalchosun.dizzo.com)