Chile launched Latam-GPT in February 2026, a regional initiative seeking to establish Latin America’s own large language AI model as a move toward greater technological sovereignty. The project, supported by 65 institutions across 15 countries, aims to build a culturally relevant AI system addressing regional languages and nuances, while reducing dependence on dominant US and Chinese tech firms.
What happened
Led by Chile’s National Center for Artificial Intelligence (CENIA), Latam-GPT received funding from the Development Bank of Latin America and the Caribbean (CAF), the Inter-American Development Bank (IDB), and Chile’s Ministry of Science. Described at launch as “the first model developed by and for Latin America,” the initiative focuses on correcting regional representation gaps and enhancing local AI capabilities.
Latam-GPT is built upon Meta’s open-source LLaMA model through continued pre-training, using Amazon Web Services infrastructure. The project also developed Copuchat, a public chatbot interface based on OpenAI’s API, to gather training data. These facts were not emphasized at launch, leading to some criticism over transparency regarding the project’s true autonomy and reliance on foreign technology.
The consortium approach involves government, academia, and civil society aiming to position Latin America as an active player in global AI governance, rather than merely consumers of externally developed AI tools.
Why it matters
The Latam-GPT initiative highlights the complexities of achieving AI sovereignty in regions with limited AI infrastructure and funding. Latin America’s aspiration to reduce dependency on US and Chinese technologies confronts practical challenges, such as the costs of building AI models from scratch and maintaining sustainable funding and collaboration across countries.
Questions persist about what sovereignty means in practice: whether Latin America should own, control, or govern entire parts of the AI development stack, or selectively partner and share resources. The unclear goals and limited disclosure about reliance on foreign AI architectures complicate assessing the initiative’s long-term impact on regional autonomy and competitiveness.
Moreover, the ability of regional actors to use the generated datasets effectively depends on infrastructure and expertise that may remain constrained. These issues underscore the importance of clarifying sovereignty objectives and engaging in transparent, inclusive regional dialogue.
Background
The concept of AI sovereignty varies widely, encompassing control over data, infrastructure, and AI models. Latin America’s effort reflects broader global trends where countries like India and Singapore develop national AI models, while others focus on managing dependencies strategically through partnerships. Canada’s Sovereign AI Compute Strategy, for instance, explicitly defines locally controlled AI components, illustrating varied approaches.
Latam-GPT’s emphasis on collaboration across multiple sectors and countries represents a regional push toward AI autonomy. However, similar initiatives often face sustainability challenges if funding and institutional support wane. The Latin American consortium could alternatively leverage its curated data corpus to influence global AI companies to incorporate regional contexts, potentially providing a more feasible form of sovereignty.
Clear definitions of AI sovereignty and strategic priorities remain crucial for Latin America to shape policies, investments, and regulations that align with its technological and cultural needs.
Sources
This article is based on reporting and publicly available information from the following source:
Read more World News stories on Goka World News.