Mark Zuckerberg bets on voice controls to make the metaverse more intuitive

The metaverse, that virtual world proposed by Meta (Facebook's matrix), continues to cook.

Its fundamental ingredient is artificial intelligence (AI), the technology that will make it possible for everything to work.

The CEO of Meta, Mark Zuckerberg, has shown this afternoon at a corporate event some specific projects that his team of AI researchers is working on and that, in his opinion, will be essential for "the immersive internet" to launch walk.

They all have one element in common: the voice.

Among the challenges that the company faces for users to navigate this new reality is the creation of a new generation of digital assistants.

In the metaverse, which we will access with virtual reality glasses, we will receive a large number of visual stimuli in the first person.

In order not to be buried in the face of so much information, it will be essential to improve the interaction between the machine (the metaverse) and the users (the avatars).

The easiest way to have fluid communication is to be able to converse with the system, with a voice assistant that learns from us.

Although it has no official name, it is referred to as the CAIRaoke Project.

A video showed the potential of this tool if it is combined with mixed reality glasses (those that superimpose digital images on physical reality).

The example: a man cooks in his house and a voice gives him the instructions of the steps to follow as the ingredients that he has in front of him light up for him to use.

01:29

This is how BuilderBot works

Mark Zuckerberg showed in this video how BuilderBot works, the scenario construction assistant in the metaverse.

Zuckerberg has also shown BuilderBot during his presentation, a kind of generator of virtual worlds to which spoken instructions are given.

“Put in a sea,” says Zuckerberg's avatar on a grid, and suddenly a digital sea materializes.

“Now put an island and a sky with clouds”, orders the tycoon so that they appear next.

a universal translator

More ambitious than these assistants is the multilingual translator in real time that Meta prepares.

Until now, existing translators convert any text or audio in an input language into English and then from English it is translated into the output language.

Thus, for a translation from Russian to Spanish, the machine would digitize the Russian text or audio and transform it into characters expressed in English, which would then undergo the reverse process and be converted into Spanish.

This method has proven to be effective, but it increases the possibility of errors and missing nuances.

models will be used

so that the translation is straightforward.

The goal is to make encounters between avatars of different nationalities in the future metaverse as pleasant as possible.

"The ability to communicate with anyone in any language is a superpower we've always dreamed of," Zuckerberg said.

In his metaverse, according to his plan, we will all have that ability thanks, once again, to AI.

The company's goal is to include the 20% of the world's population (about 1.5 billion people) who do not speak any of the world's most popular languages, such as English, Spanish or Mandarin, and who are therefore less supported by conventional translation tools.

Zuckerberg has referred to the new Meta supercomputer dedicated to AI research, which will help make the detailed projects possible.

But he didn't mention anything about the Quest VR goggles, the gateway to the metaverse, which are rumored to be ready to go into production.

“The types of experiences that the metaverse will offer go beyond what is possible today in the digital sphere.

You're going to be able to feel what you see and touch, and that's going to require advances in a wide range of areas, both hardware and software.

The key to unlocking these advances is AI”, Zuckerberg stressed during his speech.

You can follow