How Latam-GPT Will Empower Latin America How Latam-GPT Will Empower Latin America The National Center for Artificial Intelligence (CENIA) in Chile is leading the development of a large language model (LLM) for Latin America known as Latam-GPT. The new model is expected to launch by the end of 2025. Latam-GPT has been in development since 2023. As of February 2025, it was capable of processing at a capacity comparable to OpenAI’s ChatGPT-3.5. The project is open-source and free to use, capable of communicating in Spanish, Portuguese and several Indigenous languages. Latam-GPT has the potential to empower underprivileged people in Latin America by expanding access to artificial intelligence (AI) tools and education.

Why Develop an LLM for Latin America?

Álvaro Soto, director of CENIA, explained why context matters in an interview with WIRED: “Imagine if we wanted to use them to modernize the education system in Latin America. If you ask one of these models for an example, it would probably tell you about George Washington.”

According to the Economic Commission for Latin America and the Caribbean (ECLAC), digital skills in the region lag behind those in wealthier countries. Only about 30% of adults had basic digital skills in 2020, compared with 80% in developed nations.

There is also a growing divide between people who work in jobs with access to AI tools and those who do not. Of the 30% to 40% of jobs that use generative artificial intelligence (Gen AI) in Latin America, most are concentrated in urban, higher-paying sectors that require advanced education. Meanwhile, women are at greater risk of being replaced by AI automation.

A Joint Effort

To build an LLM for Latin America, CENIA is collaborating with organizations from Mexico, Argentina, Colombia, the United States, Spain, Peru and Uruguay. As of October 2025, its computing center in Arca, Chile, had processed 2,645,500 documents from Central and South America, Spain and the Caribbean. Progress depends on each country’s level of digital development and the availability of structured information. Latam-GPT also trains its model using conversations from users through Copuchat, an online chat platform that allows anonymous interaction and information sharing. Copuchat currently runs on GPT-3.5.

The Role of Latam-GPT in Reducing Poverty

An open-source LLM that understands the languages and cultures of Latin America can help reduce inequality and foster inclusion in several ways:

  • Low-cost AI tutors. CENIA’s collaboration with Mapuche, Rapanui and Guaraní translators ensures the preservation of Indigenous languages while enabling access to educational tools.
  • Upskilling the workforce. Latam-GPT can provide affordable training and assessments for employees and employers, expanding opportunities for professional growth.
  • Improving health education. In many Latin American cultures, people often underestimate the seriousness of their symptoms. Latam-GPT could provide a platform for discussing health concerns and encouraging individuals to seek medical care when needed.
  • Supporting startups, researchers and NGOs. Many organizations lack the resources to purchase expensive AI software. A Microsoft Copilot license costs about $30 per user in addition to a 365 subscription, and companies have spent an average of $400,000 annually on AI-native applications in 2025. Latam-GPT’s open-access model offers a cost-effective alternative.
  • Reducing cultural bias. Indigenous and Latin American cultures are often underrepresented in models trained primarily in English. Latam-GPT seeks to correct this imbalance by incorporating regional data and perspectives.
  • Empowering small businesses: Access to AI tools and market information can help small businesses become more competitive. A Deloitte study found that digital tools can make small businesses twice as profitable.

Looking Forward

Through CENIA’s efforts, Latam-GPT aims to empower Latin America to learn and work at a competitive level regardless of socioeconomic status. A model trained in Spanish, Portuguese and more than 50 Indigenous languages will allow for broader inclusion and the preservation of cultural identity. The future of Gen AI as an essential tool is inevitable. Latam-GPT can help bridge the digital and educational divide, allowing even the most poverty-stricken regions to access information, training and technology that might otherwise remain out of reach.

– Johanna Lorena Arredondo Gonzalez

Johanna is based in Pittsburgh, PA, USA and focuses on Technology and Global Health for The Borgen Project.

Photo: Flickr

Source link