Mexico City (AFP) – Artificial intelligence companies are training a language model from indigenous languages spoken in Mexico, the developers behind the project announced Friday.
Published on: 06/11/2026 - 20:10
“What we are looking for is to ensure that communities are not left behind,” Jessica Peña, founder of Monarca Lab, told AFP.
Some 68 indigenous languages are recorded in Mexico. But their speakers generally live in poor localities, without access to technology.
A model adapted to indigenous languages is nourished by texts, audio recordings, translations and grammars specific to these languages. This allows the AI to learn the structure, vocabulary and nuances of these traditional dialects.
The language model is in the testing phase and its creators hope to have results within a year, Jessica Peña added.
Brazilian company WideLabs, co-creator of the initiative, has already developed a similar language model in Portuguese, which takes into account the language variants in that country.
This model has been implemented in Brazil by justice institutions, in applications that direct citizens to access legal services.
The Monarca Lab announcement took place as part of Mexico IA+, a meeting between specialists in the artificial intelligence sector.
The government of President Claudia Sheinbaum is seeking to boost the development of AI as part of Plan Mexico, the aim of which is to replace imports amid trade tensions with the United States.
Transportation of electric power and access to capital pose challenges for the sector, said Octavio Camarena, director of KIO, a Mexican data center company.
However, Mexico, Latin America's second-largest economy, continues to attract foreign companies.
Illumia, a company of Argentinian origin, arrived there about three weeks ago due to the “large number of companies” that need AI services, said its representative in the country, Oscar Parra.
© 2026 AFP
The content you are trying to access does not exist or is no longer available.
