Facebook owner Meta has announced a major advance in its AI language translation technology as the company seeks to showcase its translation credentials amid criticism of poor content moderation in less common languages.
Meta’s AI model, which is part of its No Language Left Behind project, has the capacity to translate 200 different languages across 50 billion parameters using one of the world’s fastest AI supercomputers, and is already used to translate content on Facebook and Wikipedia.
The technology covers almost double the number of languages offered by conventional translation software with improved accuracy, and will enable over 25 billion translations to be completed by AI every day across Meta’s apps.
Meta founder Mark Zuckerberg said in a Facebook post: “Communicating across languages is one superpower that AI provides, but as we keep advancing our AI work it’s improving everything we do — from showing the most interesting content on Facebook and Instagram, to recommending more relevant ads, to keeping our services safe for everyone.”
It comes after Meta was accused of failing to moderate harmful content in different languages following an investigation by human rights group Global Witness which found the platform approved adverts containing violent hate speech written in Amharic, a language common in in Ethiopia.
In a statement to the Guardian Meta said: “We’ve invested heavily in safety measures in Ethiopia, adding more staff with local expertise and building our capacity to catch hateful and inflammatory content in the most widely spoken languages, including Amharic.”