AI software company Konan Technology unveiled its on-device AI technology, which integrates its self-developed STT, TTS, and LLM into smartphones.
On the 14th, Konan Technology posted a video titled "Konan On-Device" on its official YouTube channel. The Konan On-Device was demonstrated on Samsung Electronics' Galaxy S24 Ultra model. When the microphone button is tapped and spoken to in Korean, the device automatically recognizes the speech (Speech-to-text), translates it in real time, and provides the translated speech in English (Text-to-speech) to the other party.
The system translates what the user says into the other party’s language and plays it back as speech, while displaying the translation results of both the user’s and the other party’s languages in a chat window on the smartphone screen. In addition to Korean and English, more supported languages will be added through updates.
Although voice recognition was already possible with existing ChatGPT, Konan Technology added voice synthesis functionality. The process of recognizing spoken words on the device without network connection, translating them into the other party’s language, and speaking them is all handled seamlessly by AI. Since it uses only the mobile device’s CPU, it can be efficiently utilized on various edge devices at a low cost.
Konan Technology highlighted that its core competitive edge in voice technology lies in its neural network design specialized for high-speed real-time processing, enabling recognition results to be output immediately without response delay (STT) and synthesized instantly (TTS).
Kim Youngseom, CEO of Konan Technology, said, "We are consolidating generative AI technology capabilities to implement optimized on-device AI technology for each terminal," adding, "We will create diverse demand channels and business opportunities and convert them into sales."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
