Chile launches Latam-GPT to address Latin America’s AI blind spots

Chile-backed Latam-GPT, a regionally-focused AI model, plans to make its public debut in September after completing initial trials. The project aims to address gaps in AI representation for Latin America, where existing models often provide incomplete or stereotypical information about the region, while positioning Latin American countries as active participants rather than passive consumers in the global AI race.

What you should know: Latam-GPT is being developed by Chile’s National Center for Artificial Intelligence (CENIA), a government-funded research center, with backing from multiple Latin American governments and institutions.

• At least 37 entities—including government agencies, universities, the Organization of American States, and Inter-American Development Bank—have agreed to contribute data, with 70 more in negotiations.
• The model will be trained primarily on Spanish and Portuguese data, contrasting with English-dominant models like ChatGPT.
• Unlike many major AI models that don’t fully disclose their data sources, Latam-GPT will publish its complete list of training sources.

Why this matters: The project represents Latin America’s effort to build technological sovereignty and ensure accurate AI representation of the region.

• “If we want to be part of the debate on governance, adoption, and putting AI at the service of people in our countries … we have to get our hands dirty,” said Rodrigo Durán, CENIA’s director.
• Current AI models trained elsewhere often give incomplete or overly stereotypical answers about Latin America due to limited regional data.

How it works: CENIA is implementing ethical data curation as a core principle, anonymizing sensitive personal information before training.

• Engineers are carefully reviewing all contributed data to protect privacy while maintaining utility.
• The transparent approach to data sourcing has encouraged widespread institutional participation.

Potential applications: Several Chilean city governments are exploring practical uses through workshops focused on public service improvements.

• Applications include reducing school dropout rates and decreasing wait times for public health services.
• Chile’s Gabriela Mistral University expects to enhance AI-powered educational tools, such as making virtual patients behave in more culturally appropriate Chilean ways rather than defaulting to American or European behaviors.

Key limitations: Funding constraints will affect the model’s deployment and capabilities.

• Due to tight budgets, the team plans to release Latam-GPT as downloadable code rather than a conversational chat application like ChatGPT.
• Limited computing power means the model will be weaker at complex mathematical tasks but stronger in social sciences and humanities.

The big picture: The initiative echoes other geographically-focused AI models like Singapore’s Sea-Lion, which trains on 13 Southeast Asian languages.

• Fernando Vargas of the Inter-American Development Bank sees AI cooperation as a potential bright spot for Latin American integration, noting that participants could “develop models in other areas of knowledge.”
• However, economist Eduardo Levy Yeyati warns that basic infrastructure gaps in internet reliability and affordability could limit AI’s regional potential, stating: “We have public education, health, and transportation policies, but we don’t have public digital connectivity policies.”