Meta plans to launch an AI-integrated universal speech translator in the future / Digital Information World

Mark Zuckerberg is working on a Metaverse project that aims to create AI-based software for language translations around the world.

Even though access to technology is growing globally, nearly half of the world’s population still cannot access online platforms in their language, as most AI software used for translations today have often been educated in the international language, English. Thus, many services translate something from a source language to English and then from English to the destination language, which adds noise and inaccuracy to the translation. Meta’s CEO plans to create an open-source AI model that can directly translate between hundreds of languages without having to use English as an intermediate language.

Meta claims to launch two new projects. First, they want to develop a model that leaves no room for language. Every language will be translated whether it is full of text or not. The artificial intelligence system will decode all languages, from Korean to Urdu or from Japanese to Hindi, which will facilitate communication and understanding of information on the Internet. Second, Meta planned to create an ability to speak with others in your native language.

With these very advanced projects, Meta is eager to deliver better translations, allowing people to socialize more. To date, Meta is trying to innovate in self-supervised learning and launching supercomputers to give deep insight into open science. The Metaverse project is a game changer in many ways as it is designed to provide a virtual world where you can communicate and interact with objects in space. This AI-powered language translator is reprogrammed on machine learning. These digital machine learning tools will allow the system to translate any language, so that one can share information among others without worrying.

Read next: English isn’t as popular as you think and many languages ​​are underrepresented on the internet