Latin America to Launch Latam-GPT: First Regional AI Language Model Set for September Release
In a landmark move to democratise artificial intelligence across Latin America, a coalition of over a dozen countries is set to launch Latam-GPT, the first large-scale AI language model trained specifically for Latin American languages and cultures.
The open-source project is spearheaded by Chile’s National Centre for Artificial Intelligence (CENIA) in collaboration with more than 30 regional institutions and is expected to go live by September 2025, Chilean officials announced on Tuesday.
A Regional AI Built for Latin America’s Cultural and Linguistic Diversity
Unlike global models trained mainly on English content, Latam-GPT was designed to reflect Latin America’s cultural context, regional dialects, and Indigenous languages, thereby improving accuracy and inclusiveness in AI applications.
“This model could be a democratizing element for AI,” said Chilean Science Minister Aisen Etcheverry, highlighting its potential use in schools, hospitals, and digital public services.
Latam-GPT is built on Meta’s Llama 3 architecture and trained using a distributed network of computing systems, including on-premise infrastructure at the University of Tarapacá in Chile and cloud computing from Amazon Web Services (AWS).
Key Features of Latam-GPT:
- Open-source and locally trained
- Built with cultural and linguistic sensitivity
- Initial Indigenous language support (e.g., Rapa Nui)
- Designed for public sector and educational use cases
- Powered by Llama 3 and supported by regional infrastructure
Preserving Indigenous Languages Through AI
One of Latam-GPT’s standout missions is to preserve Indigenous languages. Developers have already created a translator for Rapa Nui, the native language of Easter Island, and plan to extend support to other under-represented languages throughout the region.
These tools will be used in creating AI-powered virtual assistants, multilingual education platforms, and public service applications that are accessible in both major and Indigenous languages.
Not a ChatGPT Rival, But a Core Regional Tech Platform
Officials emphasise that Latam-GPT is not designed to compete with commercial AI tools like ChatGPT but to serve as the technological backbone for locally relevant applications.
“We’re not trying to build another chatbot,” said CENIA Director Alvaro Soto. “We’re creating the foundation for a localized AI ecosystem that can power innovation in public services, education, and beyond.”
Despite lacking a dedicated budget, the project has received support from the CAF, Latin American Development Bank, as well as Amazon Web Services. Soto remains optimistic that showcasing Latam-GPT’s real-world applications will attract more investment.
Why Latam-GPT Matters:
- First AI model tailored for Latin America
- Promotes inclusion of non-English languages in AI
- Supports education and public service digital transformation
- Fuels local innovation through open-source collaboration
- Encourages AI adoption with cultural relevance and accessibility
