
Juan N. Pava
Research - Post-Bacc, Ethics In Society
Bio
Juan N. Pava is a Research Fellow in the Tech Ethics and Policy Rising Scholars Program at Stanford’s McCoy Family Center for Ethics in Society. At Stanford HAI, his work focuses on the intersection of emerging technologies, the social sector, and the Global South, with an emphasis on equity and access. Separately, he collaborates with Stanford’s Human-Trafficking Data Lab, where he investigates issues of labor exploitation.
Juan’s broader research interests include the political economy of emerging countries and its intersection with political philosophy and ethics. He holds a B.A. in Philosophy and Economics from New York University and was born and raised in Colombia.
All Publications
-
Mind the (Language) Gap: Mind the (Language) Gap: Mapping the Challenges of LLM Development in Low-Resource Language Contexts
Stanford Institute for Human-Centered AI.
2025
Abstract
As large language models (LLMs) proliferate, they increasingly shape who is empowered by AI—and who is left behind. This paper maps the technical and ethical challenges of LLM development for low-resource lan-guages, which encompass over 95% of the world’s linguistic diversity yet remain marginalized in mainstream model development. We analyze three key technical approaches to addressing this divide—massively multilingual, regional multilingual, and monolingual models—and assess their trade-offs in terms of accuracy, scalability, and cultural representation. We then examine approaches to close the underlying data gap, including ma-chine translation pipelines, semi-automated strategies, and community-driven dataset creation. We ultimately argue that addressing the digital language divide is critical to ensuring that the benefits of AI are equitably distributed—and that its harms are not disproportionately borne. By offering this landscape-level overview of cur-rent efforts across model design and data creation, we aim to draw attention to this issue and provide a foundation upon which further research can build. We conclude with three high-level recommendations to AI researchers, funders, policymakers, and civil society organizations: in-crease investment in low-resource language infrastructure and collaboration; prioritize AI development initiatives that are community-driven and culturally grounded; and ensure fair practices in the collection, access, and owner-ship of language data.
-
Response to USAID's Request for Information on AI in Global Development Playbook
Stanford Institute for Human-Centered AI.
2024
Abstract
In this response to the U.S. Agency for International Development’s (USAID) request for information on the development of an AI in Global Development Playbook, scholars from Stanford HAI and The Asia Foundation call for an approach to AI in global development that is grounded in local perspectives and tailored to the specific circumstances of Global Majority countries.
-
Inspiring Action: Identifying the Social Sector AI Opportunity Gap
Stanford Institute for Human-Centered AI.
2024
Abstract
This national survey was a collaboration between Stanford’s Institute for Human-Centered Artificial Intelligence and Project Evident and was conceived as a project to shed light on the current use of, interest in, and opportunity for AI in the social and education sectors. Over the last decade, AI has reshaped the commercial sector and consumer habits, resulting in significant value creation and profitability—think value created by recommendation systems in e-commerce or streaming services. As it becomes easier to include AI applications (Microsoft Copilot, Google Workspace, OpenAI GPTs) as part of the nonprofit technology stack, the social and education sectors have the same opportunity to deploy AI to create value through enhanced mission-related outcomes.