Transforming Medical Language Recognition: AI Solutions by PuzzleAI

Innovative AI Solutions Transform Medical Language Recognition

Medical professionals in South Korea often mix medical terminology in Korean and English, creating a language that resembles an ‘alien language.’ To address this issue, we have enhanced our AI voice recognition technology to an unprecedented level, making it competitive enough to attract collaboration proposals from big tech companies,” stated Kim Yong-sik, CEO of PuzzleAI, in an interview with the Seoul Economic Daily on the 24th.

PuzzleAI has developed the AI-based voice recognition electronic medical record system, called VoiceEMR, and the voice recognition nursing record system, VoiceENR. According to the company, the voice recognition accuracy of VoiceEMR is approximately 99.7%. Currently, 172 medical institutions in South Korea have adopted PuzzleAI’s solutions.

Expanding into Global Markets

The company has developed voice recognition technology capable of processing complicated medical terms even when mixed with unclear pronunciations in Korean and English, inadvertently boosting its global competitiveness. Kim noted, “Recently, hospitals in the eastern United States showcased our VoiceEMR system, receiving positive feedback. However, recognizing voices of speakers whose native language is English is relatively easy. We train our system to recognize voices of non-native English speakers, allowing accurate recognition even when users from countries like China, India, and Saudi Arabia use imprecise English. This provides significant potential for market expansion overseas.”

PuzzleAI is now leveraging its outstanding voice recognition technology to expand into generative AI. At the Global Health Expo (GHE) held in Saudi Arabia from the 22nd to the 24th, the company debuted a specialized generative documentation solution that utilizes a Large Language Model (LLM) combined with medical data, enabling it to generate patient-doctor conversation charts in real-time.

A Bright Future with Generative AI

The rise of generative AI presents new opportunities for PuzzleAI. “To successfully launch generative AI projects, we need robust voice recognition technology at the forefront, which has led to numerous collaboration requests from both domestic and overseas corporations facing challenges in this area. Compared to medical voice recognition, the voice recognition requirements in various other industries are not as demanding, allowing PuzzleAI to tackle complex and challenging voice recognition issues effectively,” Kim added.

Furthermore, PuzzleAI is developing a medical imaging generation solution that predicts patients’ visual fields using Optical Coherence Tomography (OCT) images. This solution recently received an award at the American Academy of Ophthalmology (AAO) meeting held from the 19th to the 21st, recognizing its technological capabilities. Kim concluded, “We initiated this development based on proposals from domestic ophthalmologists, and our company has been pleasantly surprised by the potential of our technology. We anticipate a range of collaboration opportunities in the AI and voice recognition fields moving forward.

Original article written by Park Hyo-jeong for Seoul Economic Daily