Chile unveils Latam-GPT, an open-source AI model built to combat bias in Latin America

Chile Launches First Open-Source AI Model for Latin America

Chile unveiled Latam-GPT on Tuesday, marking a significant step toward regional technological independence in artificial intelligence. The open-source AI model represents a two-year collaborative effort led by the Chilean National Center for Artificial Intelligence (CENIA) and supported by over 30 institutions across eight Latin American countries including Argentina, Brazil, Colombia, Ecuador, Mexico, Peru, and Uruguay1

1

. President Gabriel Boric emphasized the strategic importance of the launch, stating that "artificial intelligence is the greatest technological revolution of recent times, and from Latin America and the Caribbean, it is strategic and urgent that we play a role"1

1

. The initiative aims to position the region as an active player rather than a passive recipient in the global AI economy.

Addressing Linguistic Bias Through Regional Data Training

The artificial intelligence language model was developed specifically to combat US-centric bias inherent in existing AI systems trained primarily on English data. Rodrigo Durán, executive director at CENIA, explained that "Latam-GPT is trained with a proportion of Latin American data that previously did not exist online and was not included in existing models"1

1

. The model required collecting more than eight terabytes of data, equivalent to millions of books, from private sources obtained through strategic partnerships and synthetic data addressing underrepresented areas1

1

. This comprehensive data training enables the model to recognize regional slang, idioms, and even speech patterns that other AI systems often misinterpret3

3

.

Foundational AI Infrastructure for Regional Applications

Unlike consumer-facing tools such as ChatGPT or Google Gemini, Latam-GPT functions as foundational infrastructure for future regional AI applications rather than a direct competitor1

1

. The open-source nature allows programmers to customize the software to suit specific local needs3

3

. CENIA director Alvaro Soto highlighted potential applications in healthcare, particularly for hospitals "with logistical problems or issues with the use of medical resources"3

3

. Chilean company Digevo has already committed to using the model to develop conversational robots specializing in customer service for airlines and retail companies, with clients showing strong interest in having users express themselves in local language2

2

.

Building Technical Capacity on a Modest Budget

The AI model for Latin America was developed with just $550,000 in funding, primarily sourced from the Development Bank of Latin America (CAF) and CENIA's own resources1

1

. The first version was developed using Amazon Web Services cloud infrastructure and will launch at the end of February4

4

. Future versions will be trained on a supercomputer at the University of Tarapacá in northern Chile, with an investment of approximately $4.5 million, starting in the first semester of 20261

1

. While academics like Alejandro Barros from the University of Chile acknowledge that Latam-GPT has "no chance" of competing with major global AI models due to resource disparities, the project demonstrates that Latin America can develop and understand this technology3

3

.

Preserving Cultural Diversity and Language Representation

The project operates primarily in Spanish and Portuguese, with plans to incorporate indigenous languages in later stages1

1

. Gabriela Arriagada, a researcher at CENIA and head of the project's ethics team, explained that "when we talk about incorporating Latin American culture, we are referring to a training approach designed to address data that reflects cultural realities, identifying where gaps exist in other models"1

1

. Chile's Science Minister Aldo Valle emphasized that Latin America "cannot simply be a passive user or recipient of artificial intelligence systems" as this could result in the loss of significant traditions3

3

. The initiative joins other regional AI efforts worldwide, including SEA-LION in Southeast Asia and UlizaLlama in Africa, all focusing on their own cultural contexts2

2

.

Implications for AI Regulation and Future Development

The development of Latam-GPT carries important implications for AI regulation in the region. Durán noted that "you cannot regulate something you do not understand," and the project demonstrates that Latin America now has the technical capacity to build AI models1

1

. Luis Chiruzzo, an engineering professor at the University of the Republic in Uruguay, called it "a very important milestone for Latin America" that provides assurance that everyone is included in the training1

1

. The initiative comes as the United States, China and the European Union control more than half the world's most powerful data centers, while Africa and South America have almost no AI hubs according to Oxford University data4

4

. The project was announced at the February 2025 Artificial Intelligence Action Summit in Paris and launched in early 20231

1

.