Meta Partners with UNESCO to Improve AI Speech and Translation for Diverse Languages

Meta has announced a partnership with UNESCO to launch the Language Technology Partner Program, an initiative designed to collect speech recordings, transcriptions, and translated texts in a wide range of languages. The goal is to develop open-source AI models that enhance speech recognition and translation technologies, with a particular focus on underserved languages.

The program is seeking collaborators who can contribute over 10 hours of speech recordings with transcriptions, large volumes of written text, and translated sentence sets in diverse languages. Participants will work closely with Meta’s AI teams to integrate these contributions into future models. To participate, please complete this form.

Focus on Underserved Languages

One of the program’s first partners is the government of Nunavut in Northern Canada, where residents speak Inuktut, a collection of Inuit languages. This collaboration reflects Meta’s broader commitment to addressing languages that are often overlooked in mainstream AI development.

“Our efforts are especially focused on underserved languages, in support of UNESCO’s work,” Meta stated in a blog post shared with TechCrunch. “Ultimately, our goal is to create intelligent systems that can understand and respond to complex human needs, regardless of language or cultural background.”

Complementary AI Tools and Benchmarks

In addition to the Language Technology Partner Program, Meta is releasing an open-source machine translation benchmark. This benchmark, crafted by linguists, evaluates the performance of translation models across seven languages and is available on the AI platform Hugging Face. Developers and researchers can access—and contribute to—this resource to improve language translation models collaboratively.

While Meta frames these initiatives as philanthropic, they also serve the company’s interests by enhancing its AI tools, including Meta AI, its AI-powered assistant, and features like automatic translation for creators.

AI Development and Past Criticisms

Meta has faced significant criticism for its handling of non-English content across platforms like Facebook and Instagram. Reports highlight issues such as:

• 70% of COVID misinformation in Italian and Spanish was left largely unflagged, compared to 29% English-language content.

• Arabic-language posts being mistakenly flagged as hate speech, according to leaked internal documents.

In response, Meta claims it is actively improving both its translation and moderation technologies. For instance, last September, Meta began testing a tool to translate voices in Instagram Reels, allowing creators to dub their speech and auto-lip-sync it for multilingual audiences.

Looking Ahead: A Step Toward Inclusive AI?

Meta’s partnership with UNESCO and its focus on underserved languages marks a significant step toward building more inclusive AI systems. By open-sourcing these models and inviting global contributions, the company has the potential to democratize access to advanced speech and translation technologies.

However, Meta’s track record with content moderation and its commercial interests in AI raise questions about how these tools will be implemented and who will ultimately benefit. As Meta continues to expand the language capabilities of its AI products, transparency and accountability will be critical in ensuring that these advancements genuinely support diverse linguistic communities around the world.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.