What is Latam-GPT: AI Powering Spanish and Portuguese in Latin America?

A New Initiative for Latin American AI
Chile has launched an ambitious project called Latam-GPT, which aims to provide Latin America with its own artificial intelligence (AI) model. This initiative is designed to address the dominance of American companies in the AI sector and reduce the biases present in current systems. The project is being led by Chile's National Centre for Artificial Intelligence (Cenia), a private organization that receives public funding. It has also received support from various institutions across several Latin American countries, including universities, foundations, libraries, government entities, and civil society organizations.
During the official launch of the project on Television Nacional, Chilean President Gabriel Boric emphasized the region’s potential in global technological development. He welcomed the opportunity for Latin America to become a key player in the future economy. The president highlighted that Latam-GPT will help the region move beyond being a passive consumer of AI technologies developed elsewhere, which could lead to the erosion of cultural traditions.
Understanding Latam-GPT
Despite its name, Latam-GPT is not an interactive chatbot. Instead, it is a large database trained on regional information, intended as a foundation for developing technological applications tailored to the needs of Latin America. The model was created using over eight terabytes of data—equivalent to millions of books—and funded with $550,000 (approximately €460,000). This funding came primarily from the Development Bank of Latin America (CAF), along with contributions from other institutions.
The first version of the system was developed on Amazon Web Services. In the future, the model will be trained on a supercomputer to be installed at the University of Tarapacá in northern Chile by mid-2026, with an investment of nearly five million dollars.
Regional AI Initiatives
The development of large AI models has traditionally been concentrated in the United States, China, and Europe. However, other regional initiatives have emerged, such as SEA-LION in Southeast Asia and UlizaLlama in Africa, each focusing on their unique cultural contexts. Alvaro Soto, director of Cenia, noted that models developed elsewhere often include only a small proportion of Latin American data. President Boric illustrated this point by comparing the extensive information available on the siege of Calais with the limited coverage of key battles in Chilean history, such as the siege of Chillán.
At this stage, the content of Latam-GPT is mainly in Spanish and Portuguese. However, the goal is to incorporate indigenous languages in the future. Despite these efforts, some experts, like Alejandro Barros from the University of Chile, argue that Latam-GPT may struggle to compete with global AI models due to differences in economic resources and infrastructure.
Potential Applications
Latam-GPT will be a free platform, offering tools that can be adapted to local needs. For example, it could be used to develop digital solutions for hospitals facing logistical challenges or for optimizing the use of medical resources. One of the first companies to use the model will be Digevo, a Chilean firm that plans to create conversational robots for customer service in airlines and retail. According to Digevo’s director, Roberto Musso, companies are interested in having users interact in their native language.
Musso also mentioned that Latam-GPT can recognize slang, idioms, and even speech patterns, which helps reduce bias issues commonly found in other AI models. This feature makes it particularly suitable for local markets where linguistic diversity is a key factor.
Future Prospects
As the project moves forward, the focus will be on expanding the model’s capabilities and ensuring it meets the diverse needs of Latin American users. With continued support from various stakeholders, Latam-GPT has the potential to play a significant role in shaping the region’s technological landscape and promoting digital sovereignty.