META announces plans to build an AI-powered 'universal speech translator

Meta, the owner of Facebook, Instagram and WhatsApp, has announced an ambitious new AI research project to create translation software that works "for everyone in the world." The project was announced as part of a program focused on the broader benefits of Meta that AI believes can offer the company's Metaverse plans.

“The ability to communicate with anyone – in any language – is something people have dreamed of forever, and AI is going to accomplish that in our lifetime,” Mark Zuckerberg, Meta CEO, said in an online presentation.

The company says that although commonly spoken languages such as English, Mandarin and Spanish are well served by current translation tools, about 20 percent of the world's population does not speak the languages covered by these systems. Often, these under-served languages do not have easily accessible collections of written text that are needed to train AI systems or sometimes no standardized writing systems.

Meta says it wants to overcome these challenges by deploying new machine learning techniques in two specific areas. The first focus, called No Language Left Behind, will focus on building an AI model that can learn to translate language using fewer training examples. The second, Universal Speech Translator, will aim to create systems that directly translate speech from one language to another in real time, without the need for a written component to act as an intermediary (a common technique for many translation apps). translates.

In a blog post announcing the news, the META researchers did not offer deadlines for completing these projects or even a roadmap for major milestones in reaching their goals. Instead, the company emphasized the utopian possibilities of universal language translation.

"Removing language barriers will be profound, making it possible for billions of people to access information online in their native or preferred language," they write. “Advances in [machine translation] will not only help people who do not speak one of the languages that dominate the Internet today; they will also fundamentally change the way people connect and share ideas in the world. "

Importantly, Meta also envisions that such technology will immensely benefit its world-spanning products – expanding their reach and turning them into essential communication tools for millions of people. The blog post notes that Universal Translation Software will be a killer app for future wearable devices like AR Glasses (which Meta is making) and will help push boundaries in the "immersive" VR and AR reality space (which Meta is also making). will break In other words, although there may be human benefits from developing a universal translation tool, it also makes good business sense for a company like Meta.

It is certainly true that advances in machine learning in recent years have greatly improved the speed and accuracy of machine translation. Many big tech companies, from Google to Apple, now provide users with free AI translation tools used for work and travel, and undoubtedly offer countless benefits around the world. But the underlying technology has its problems, too, with critics noting that machine translation misses important nuances for human speakers, injects gender bias into its output, and is capable of throwing out those pesky, unexpected errors. Only a computer can do it. Some speakers of unusual languages even say they fear losing their speech and culture if their ability to translate words is completely controlled by big technology.

It is important to consider such errors when large platforms such as Facebook and Instagram implement such translations automatically. For example, consider a case from 2017, when a Palestinian man was arrested by Israeli police after Facebook's machine translation software mistranslated a post he had shared. The man wrote "good morning" in Arabic, but Facebook translated it as "hurt them" in English and "attack them" in Hebrew.

And while Meta has long aspired to have a global reach, the company's own products remain biased toward the countries that provide the bulk of its revenue. Internal documents published as part of the Facebook Papers revealed how the company struggles to control hate speech and abuse in languages other than English. These blind spots can have incredibly deadly consequences, such as when the company failed to tackle misinformation and hate speech in Myanmar before the Rohingya genocide. And similar cases involving questionable translations have been on Facebook's oversight board to this day.

So while a universal translator is an incredible aspiration, Meta will need to prove not only that its technology is up to the task, but that, as a company, it can apply its research fairly.