Latam-GPT is a large language model developed by the National Center for Artificial Intelligence (CENIA) in Chile, in collaboration with over thirty institutions and twelve Latin American countries. This open-source AI model aims to reflect the diverse cultures, languages, and social realities of the region. It addresses the limitations and biases of global AI models trained predominantly on English data.
Latam-GPT is trained with ethically sourced, regionally contributed data. Featuring 50B parameters, its database is comparable to OpenAI's GPT-3.5. The model focuses on Latin American content, providing more accurate responses to historical, social, and cultural inquiries specific to the region. It supports applications in education and public services, preserves linguistic diversity, and empowers local communities. Latam-GPT enhances technological sovereignty and reduces reliance on external AI solutions in Latin America.
By reflecting the unique expressions, idioms, and perspectives of Latin American communities, Latam-GPT fosters a more equitable and inclusive AI landscape for the region. This initiative unlocks tailored solutions for critical regional challenges, from improving public policy and environmental monitoring to preserving non-hegemonic languages and enhancing educational tools.
