Meta

Meta at UNGA 2024

Takeaways

  • This week, world leaders gathered for the 2024 United Nations General Assembly, where we showcased our commitment to helping accelerate the achievement of the UN’s Sustainable Development Goals through open technology.
  • In addition to announcing the recipients of the inaugural Llama Impact Grants and the Llama Impact Innovation Awards, we announced four collaborations over the course of this week that show how the global community is embracing the power of open source AI. 
  • Our collaborations with UNESCO, the US State Department, Economist Impact and Stanford Deliberative Democracy Lab demonstrate the role of open source AI models in useful humanitarian programs and social progress.

This week, world leaders and international organizations gathered for the 2024 United Nations General Assembly (UNGA). We’re proud to have showcased our commitment to helping accelerate the achievement of the UN’s Sustainable Development Goals (SDGs) at UNGA this week, because we believe technology has a significant role to play in solving some of the world’s most pressing issues. 

In particular, open source AI has real potential to serve as an equalizer, providing access to the world’s most advanced models at a global scale. In addition to announcing the recipients of the inaugural Llama Impact Grants and the Llama Impact Innovation Awards, we announced four collaborations over the course of this week that show how the global community is embracing the power of open source AI. Our collaborations with UNESCO, the US State Department, Economist Impact and Stanford Deliberative Democracy Lab demonstrate the role of open source AI models in creating social progress through humanitarian programs. 

Helping UNESCO Ensure No Language Is Left Behind

On Monday, in an event with the United Nations Educational, Scientific, and Cultural Organization (UNESCO) and Hugging Face, we presented a translation interface built on Meta’s No Language Left Behind (NLLB) AI model and running on Hugging Face’s Spaces infrastructure.The model supports high-quality translation in 200 languages, including low-resourced and marginalized languages like Asturian, Luganda, Maori and Urdu. 

We created NLLB because when low resource languages are left out of important technological advancements, it puts communities at risk of being left behind. During the announcement, UNESCO emphasized the role of the interface in supporting its work on multilingualism and the International Decade of Indigenous Languages) by promoting linguistic diversity. 

Partnering With the State Department to Harness AI

We’ve also joined the Partnership for Global Inclusivity in AI, a new partnership between the US State Department and a number of leading industry voices, designed to promote safe, secure, and trustworthy AI systems that address global societal challenges. This partnership continues our work to provide global support to expand open-source AI innovation in areas across Latin America, Africa, the Middle East, and Asia. AI has incredible potential for good, and this partnership will aim to use AI to help the US, and the world, meet essential UN SDGs. 

The Benefits of Open Source AI, Backed By Data

We collaborated with Economist Impact on a white paper that was published earlier this week, assessing the benefits of open source AI. Open source AI offers an opportunity to spur innovation and increase transparency and trust in AI systems, but we need high-quality data to ensure these models fulfill their potential. Economist Impact conducted extensive research and interviews to prepare their report, and came away with valuable insights on the potential of open source AI to benefit emerging market countries in the areas of access, innovation and economic growth. 

Bringing Underrepresented Perspectives to Our AI Community Forum

We’re thrilled to announce that on October 19, we’ll be hosting our second Community Forum on Generative AI. Our first forum took place in October of last year, and brought together 1,500 participants from four countries to discuss the principles and values we want to see reflected in AI models. We believe it’s important for regions that have been historically underrepresented in discussions about innovation to be included in conversations about AI, which is why this year’s forum will include participants from India, Turkey, Saudi Arabia, Nigeria and South Africa. The forum will be held in partnership with the Stanford Deliberative Democracy Lab, and will focus on the governing principles of AI agents and chatbots. Input from the 2023 forum has been important in shaping the ongoing development of our AI products and services, and we look forward to sharing results from the 2024 forum early next year. 

Providing Grants to Encourage Impactful Uses of Open Source AI

This week, we also announced the awardees of our 2023 Llama Impact Grants and Llama Impact Innovation Awards. We launched our grants program last year to support innovative use cases of Llama 2 and Llama 3, our open source AI models, that address pressing social issues. We received over 800 applications from organizations in over 90 countries, proposing uses of Llama in education, the environment and open innovation. From that pool of applicants, we identified three leading use cases: Wadhwani AI, Digital Green and Dana Farber will use Llama 3 to improve English language learning in public schools in India, provide agricultural advice to small-scale farmers and match cancer patients to clinical trials.