Latin American countries to launch own AI model in September

In an exciting move towards advancing artificial intelligence in the region, twelve Latin American nations are joining forces to introduce Latam-GPT in September. This pioneering AI language model is specifically designed to resonate with the diverse cultures and languages found throughout Latin America, according to officials from Chile. The open-source initiative is led by Chile's National Center for Artificial Intelligence (CENIA), in collaboration with over thirty regional institutions. Chilean Science Minister Aisen Etcheverry emphasized that this project could serve as a transformative force in making AI more accessible, detailing potential applications in educational and healthcare settings that reflect local cultural contexts. Latam-GPT has been in development since January 2023 and aims to address the shortcomings of existing global AI models that predominantly focus on English. Officials clarified that rather than competing with popular consumer products like ChatGPT, Latam-GPT is intended to serve as a foundational technology for various applications, including chatbots. A significant aspect of this initiative is its commitment to preserving Indigenous languages. The team has already created a translator for Rapa Nui, the native language of Easter Island, with plans to extend this technology to other Indigenous languages, facilitating the creation of virtual public service assistants and personalized educational tools. Utilizing Llama 3 AI technology, the model is being trained across a network of regional computers, incorporating resources from Chile's University of Tarapaca and cloud systems. Support from the regional development bank CAF and Amazon Web Services has been instrumental in this effort. Although the project currently lacks a dedicated budget, CENIA head Alvaro Soto is optimistic that showcasing the model's capabilities will attract further investment.