Meta has launched a new initiative in collaboration with UNESCO to gather speech recordings and transcriptions, aiming to help develop AI systems that are accessible to a wider range of languages.
The program, called the Language Technology Partner Program, invites collaborators to contribute over 10 hours of speech recordings with transcriptions, large amounts of written text, and sets of translated sentences in diverse languages.
Meta’s goal is to work with partners to integrate these languages into AI speech recognition and translation models, which will be open-sourced once they’re developed. One of the first partners in the program is the government of Nunavut, a remote territory in Northern Canada, where residents speak Inuktut, an Inuit language.
“Our efforts are focused on supporting underserved languages in line with UNESCO’s mission,” Meta explained in a blog post. “We aim to create intelligent systems that can understand and respond to human needs, regardless of language or culture.”
In addition to the Language Technology Partner Program, Meta is releasing an open-source machine translation benchmark to evaluate the performance of translation models. This benchmark, created with linguists, supports seven languages and is available on the AI platform Hugging Face for others to contribute to and use.
While Meta presents these initiatives as philanthropic, the company stands to benefit by improving its speech recognition and translation models. Meta continues to expand the number of languages supported by its AI assistant, Meta AI, and is piloting features like automatic translation for creators. Last year, Meta started testing a feature to translate voices in Instagram Reels, allowing creators to dub their speech and sync the lip movements automatically.
Despite this effort, Meta has faced criticism for its treatment of non-English content. For example, reports have shown that Facebook failed to flag a large portion of Italian- and Spanish-language COVID misinformation and Arabic-language posts were often wrongly flagged as hate speech. Meta has stated it is working to improve its translation and moderation technologies.
