Introducing Flitto Chat Translation, Seoul’s first interactive AI translation service
Featuring advanced AI engines, this two-way screen is set to break down language barriers
This November, Flitto collaborated with the Seoul Metropolitan Government to launch the city’s first-ever interactive real-time AI translation system.
Named Flitto Chat Translation, this exciting new service uses a transparent monitor to capture a fully immersive conversational experience between people who speak different languages.
With just a few taps on the screen to choose a language, the intuitive kiosk turns spoken words into text and translates them. Its transparent screen enables real-time communication by providing a bidirectional translation that resembles chat messages.
Flitto Chat Translation in Seoul
Flitto Chat Translation is a versatile solution for any place that requires borderless communication.
Currently, you can find the smart kiosk in two major tourist information centers in Seoul: in Gwanghwamun and on the first floor of Seoul Tourism Plaza.
Flitto’s partnership with Seoul, realized through the kiosk installation, is definitely set to streamline the Seoul experience for tourists and visitors.
Unlocking a new traveling experience in Seoul
Tourists often find themselves with limited options to communicate with other people.
Conventional kiosks can provide information but offer limited flexibility for questions in different scenarios. Furthermore, human guides in information centers may not always speak the language we’re comfortable with.
Mobile translation applications are an effective solution. However, it can feel limiting to have to type, show, and relay sentences between each other. In many cases, mobile applications’ speech recognition capacities also tend to fall short.
Flitto Chat Translation focuses on preserving the face-to-face communication system between two speakers who use different languages.
The joy of communicating with locals is often an underrated traveling experience; Flitto Chat Translation hopes to amend that.
What’s the tech behind Flitto Chat Translation?
The key element that powers the smart kiosk is Flitto’s powerful AI engines.
Flitto Chat Translation operates on a system called the CT Engine, a set of engines built by Flitto. This engine is an augmented version of the existing translation and speech-to-text (STT, also called automatic speech recognition (ASR)) systems.
A great amount of high-quality language data for translation and speech recognition was added during the augmentation process. As a result, the engine’s accuracy has increased by around 20 percent.
A data-driven communication solution
The massive text and speech datasets used by Flitto Chat Translation have also been collected, processed, and refined by Flitto.
On Flitto’s platform is a massive 14-million user community consisting of language enthusiasts and bilingual users from around the world. Each user interaction within the platform is compiled and processed as Flitto DataLab’s very own non-invasive and accurate language datasets for machine learning.
Smart kiosk that gets smarter
In addition to that, Flitto Chat Translation is also built to improve itself.
Flitto’s CT Engine accumulates more speech data and translation history as people use Flitto Chat Translation. As a result, it learns to understand diverse accents and dialects more accurately and produce better translation results over time.
Supported languages
As of the first month of its release, Flitto Chat Translation supports 11 languages, including English, Chinese (Simplified), Japanese, Thai, Vietnamese, Malay, Indonesian, Arabic, Russian, Spanish, and French.
Wrapping up…
Flitto Chat Translation is a screen to break down language barriers and build the multilingual conversation experience.
Through this project in tandem with the Seoul Metropolitan Government, we are looking forward to unlock more fulfilling experiences for tourists and visitors in Seoul.
To see Flitto Chat Translation in action, you can watch the live demo here.
For more information on Flitto Chat Translation, click here.
To read more about the language data solution that made this project possible, click here.
