Meta, the creator of Facebook and the eventual metaverse, has revealed that its so-so-called Universal Translator AI is a huge success. In early demonstrations, the artificial intelligence software was capable of translating 200 different languages.
Aside from the usual languages that are in most translators, the AI supposedly knows others that aren’t always in traditional systems. While there appear to be some quality issues here and there, this could be an impressive AI from Meta.
Universal Speech Translator
This “universal speech translator” is an ambitious AI model from Meta, as the company hopes to see it grow and be across a litany of platforms. As well as social media like Facebook and Instagram, Meta is also hoping to use the AI for real-time speech in VR and AR social platforms like the metaverse.
Experts in the machine translation field talked to The Verge about this revelation, pointing out how impressive it is to see over 100 languages translated. However, the experts are also expecting less effort to be put into specific languages simply due to how many exist.
“The major contribution here is data,” Professor Alexander Fraser told The Verge. “What is significant is 100 new languages [that can be translated by Meta’s model].”
Meta AI research scientist Angela Fan, who worked on the project, said that the focus on more of these “obscure” languages is how this AI got off the ground to begin with. Despite some being worried about the lack of better-supported languages, the fact that it will support more “low-resource languages” is a good foundation to work on.
“Translation doesn’t even work for the languages we speak, so that’s why we started this project,” said Fan. “We have this inclusion motivation of like — ‘what would it take to produce translation technology that works for everybody’?”
We’re glad that it exists
Meta has faced criticism for putting all of its eggs in the metaverse basket, complete with the support of NFTs, but the existence of a universal speech translator AI is good enough for many. While it might have some imperfections, the company is open-sourcing it so someone can build upon what they’ve made and maybe even improve on it.
“Overall, I’m glad that Meta has been doing this. More of this from companies like Google, Meta, and Microsoft, all of whom have substantial work in low resource machine translation, is great for the world,” said Professor Fraser of LMU Munich. “And of course, some of the thinking behind why and how to do this is coming out of academia as well, as well as the training of most of the listed researchers.”
The aforementioned Angela Fan also brings up how the open-sourcing from Meta will be a good thing in the long run since she doesn’t expect this AI to do everything on its own. Fan even points out the limits of what Meta can do as a company since they can’t invest all of their time and money on this one AI.
“I think that’s really, really important because it’s not like one company will be able to holistically solve the problem of machine translation,” said Fan. “It’s everyone — globally — and so we’re really interested in supporting these types of community efforts.”