10 min read
Facebook, Instagram, and WhatsApp parent company Facebook announced a new AI research project that it hopes will deliver translation quality on par with professional human translation. As part of an event focused on the company’s metaverse plans, Facebook Meta announced that its AI language services group will work to translate “every language in the world” in the next five to 10 years.
“The ability to communicate with anyone in any language — that’s a superpower people have dreamed of forever, and AI is going to deliver that within our lifetimes,” said Meta CEO
According to Meta, popular languages such as English and Spanish are catered by current translation tools, about 20 percent of the world’s population do not speak in languages that are catered to by these very systems. These languages often, do not have accessible blocks of written text needed in order to train AI systems or have no standardized system altogether.
Meta says it wants to overcome these challenges for languages by deploying new machine learning techniques in two specific areas. The first focus will concentrate on building AI models that can learn to translate language using fewer training examples. The second will aim to build systems that directly translate speech in real-time from one language to another without the need for a written component to serve as an intermediary.
Meta did not announce a deadline for completing this project but instead emphasized the countless possibilities of universal language translation. They write “Eliminating language barriers would be profound, making it possible for billions of people to access information online in their native or preferred language,” further emphasizing that “Advances in [machine translation] won’t just help those people who don’t speak one of the languages that dominate the internet today; they’ll also fundamentally change the way people in the world connect and share ideas.”
Meta predicts that this technology could benefit its globe-spanning products by serving as an extension to turn them into crucial communication tools for millions of people. They stated that their universal translation software would work for future wearable devices such as AR glasses (currently being built by meta). To summarize, in addition to holding humanitarian benefits, this would serve to boost business for a company like Meta.
Recent advances in machine learning have greatly improved the accuracy and speed of machine translation. A number of big tech companies, from Google to Apple, now offer users free AI translation tools that can be used for work and tourism. But some believe that there are disadvantages to using machine translation tools. Machine Translation tools often miss nuances critical for human speakers, inject gendered bias into their outputs, and occasionally produce weird unexpected errors. Some people also fear that they may lose hold of their own speech and culture as machine translation becomes more widespread.
It is important to consider the impact of such errors when analyzing the application of such translations by platforms such as Facebook and Instagram. For example, in 2017, a Palestinian man was arrested by Israeli police after Facebook’s translation software mistranslated a post he shared by converting ‘good morning’ in Arabic to ‘hurt them’ in English and ‘attack them’ in Hebrew.
Internal documents have revealed that the company struggles to contain and moderate hate speech in languages other than English. This failure to tackle misinformation can have devastating effects, for example when the company failed to moderate hate speech in Myanmar before the Rohingya genocide. Similar cases and concerns in translations continue to plague the company’s oversight board to this day.
While universal translation is an amazing aspiration, Meta will need to ensure and prove that it can apply its research and be applicable fairly all across the board.
Top 6: Mobile App Development Frameworks....
15 min read