Overview
Recent breakthroughs in AI, and generative AI in particular, have captured the public’s imagination and demonstrated what those developing these technologies have long known – they have the potential to help people do incredible things, create a new era of economic and social opportunity, and give individuals, creators, and organizations new ways to express themselves and connect with people.
Building off a legacy of open sourcing our products and tools to benefit the global community, this past July we launched Llama 2 as the next generation of our open source large language model. Since its release, hundreds of thousands of download requests have signaled the importance of Llama 2 to the global developer community.
With open source being part of our DNA, we want to expand that opportunity to even more organizations across the globe. That is why we are launching the Llama Impact Grants program. The goal of the program is to identify and support the most compelling applications of Llama 2 for societal benefit.
We recognize that global challenges are complex and nuanced. That is why we are structuring this program around three different tracks:
Communities around the world face barriers that prevent all their members from receiving equitable access to a quality education. From supporting teachers and students with creative tooling or translation needs to personalizing and improving educational outcomes across the globe, large language models can promote opportunities that open up new troves of educational resources for those who could most benefit from them.
Applicants submitting proposals under the Education track are encouraged to consider how Llama 2 could promote the discovery of new opportunities to deliver more equitable education outcomes in their own communities. The might include (but are not limited to) the following:
- Support for Students: Deliver personalized content for individual students in order to convey key points in a way that they best understand. This could include personalized curriculum materials, course programs, and assessment questions that focus on areas where an individual child most needs support.
- Support for Teachers: Assist in grading, lesson planning, and other administrative tasks in order to allow teachers to spend more of their time teaching and supporting their students, creating a more productive teaching workforce.
Operating sustainably and addressing environmental challenges through bold, meaningful action are paramount to our mission and essential to help the world avoid the worst effects of a dynamic environment.
Applicants submitting proposals under the Environment track are encouraged to consider how Llama 2 may be incorporated as a component of a solution to build a healthier and more sustainable world, including (but not limited to) the following:
- Mitigation Focus: Help us speed the discovery, testing, and scaling of new methods for reducing the levels of green house gases in the atmosphere. Applications include accelerating material science, carbon removal, NDC strategies, and just transitions.
- Adaptation Focus: Help us adapt to a changing environment, saving lives and protecting livelihoods. Applications include agricultural planning and management; urban resilience; and green economic development.
Environment and education are only two domains where Llama 2 could be used to positively impact society. We believe there are multiple other domains, as well as intersectional ones, where our model can have an even greater positive impact on society.
We welcome all proposals that take an innovative approach to using Llama 2 for social good, including (but not limited to) the following:
- Faster synthesis and answers from large corpuses of research, policy papers, or data sets, including deriving novel insights from unstructured text
- Improving the speed of coding new software applications that help us to understand and act quickly in a changing world
- Summarizing known side effects and anecdotal outcomes of medications to mitigate negative interactions and propose solutions to increase medication compliance
- Analysis of court records and documents to identify biases and evidentiary challenges that could reopen cases and bring about more fair judicial outcomes
If your lab or organization has ideas for how to address any of these or other critical global challenges leveraging Llama 2 we welcome your proposal!
Ultimately, we hope to
- Shape Compelling Use Cases that bridge technical and non-technical communities
- Build Community that encourages adoption of open-source large language models across the globe, at local and at national levels.
- Address Global Challenges by building applications that work toward societal benefit.
Structure + Timeline:
The program will be sequenced over two phases, with an interim milestone being an in-person Refinement Workshop to provide technical support to a subset of finalist proposal teams.
October 6, 2023 - November 15, 2023
The application is now open! We encourage organizations from across the globe to submit proposals for how Llama 2 can be applied to address specific challenges aligned with one of our three tracks. Details for what to include in the proposal can be found in the program’s terms and conditions.
If interested in learning more before submitting a proposal, please review the FAQ or email LlamaImpact@meta.com. We will also be hosting two webinars on October 24th and November 7th where you can ask additional questions prior to the submission deadline.
For webinar details and registration information, please visit the links below.
October 24th webinar registration
November 7th webinar registration
The application deadline is November 15, 2023. All proposals will be evaluated using a selection criteria that assesses:
- Technical Feasibility: Llama is a central part of the proposed solution. There is a clear indication for how Llama will be used and is feasible given the capabilities of the model.
- Applicability: The proposed solution measurably addresses the challenge outlined in their designated track. The proposed solution may be novel interventions or refined through the use of Llama.
- Responsibility: Where applicable, the proposed solution articulates how it will address responsible practices with regard to data security, privacy, and replicability.
- Impact: The proposed solution articulates a vision for how their solution impacts the broader problem space and provides social benefit. This might include how the solution works across geographies, addresses the needs of additional communities, or similar.
Those teams whose proposals are selected as finalists will be notified in January 2024.
March 2024
Select representatives from each finalist team will be invited to attend an in-person refinement workshop where Meta AI researchers, open-source experts, as well as technical partners with subject matter expertise in each of the issue areas, will facilitate sessions to advise teams for how to expand their proposal into viable solutions that leverage Llama 2 to address key challenges within their sector.
March 2024 - July 2024
Following the workshop, finalist teams will be encouraged to further collaborate and continue developing their solutions. At the end of July, each finalist team will submit an expanded proposal where they provide both a technical and impact roadmap that outlines how they would implement their solution. A more detailed selection criteria along with the submission deadline will be shared at the Refinement workshop. A panel of experts will review each finalist proposal and select awardees for each track. Each team selected will receive an award of $500,000.
Who We Are
Our Actions
Newsletter
Meta © 2023