Plataforma da Diversidade Linguística Brasileira: dados linguísticos para uma IA brasileira

Generative artificial intelligence is based on large-scale language models (LLMs), which are trained with data most often collected without consent or in breach of copyright. LLMs are trained with billions of words and millions of parameters, but we don't know exactly which texts are selected in the training or which parameters are controlled. While unsupervised learning requires a large volume of data, demanding more and more computational costs and generating energy impacts, supervised learning with structured and tagged data can optimize this process; more than that: supervised learning with structured and tagged data resulting from language documentation projects can contribute directly to the National Artificial Intelligence Plan: “Develop advanced language models in Portuguese, with national data that encompasses our cultural, social and linguistic diversity, to strengthen sovereignty in AI.” In Brazil, in addition to Portuguese and its varieties, there are more than 250 other languages (indigenous, immigration, sign language), which are neglected in digital inclusion due to a lack of structured data. The consortium of laboratories and research groups in this INCT aims to prepare linguistic data for the training of LLMs, considering Brazil's linguistic diversity, with the development of a joint protocol for collecting linguistic data in the field, to be replicated in the groups and laboratories longitudinally, as well as transcription procedures, as well as procedures for transcribing, aligning and labeling linguistic data to create a data set that represents Brazilian linguistic diversity, and conducting studies on linguistic processing of diversity to fine-tune LLMs, helping to reduce asymmetries and prejudice resulting from training LLMs with translations from English.

Authors

Institutions

Publication Details

Journal
Cadernos de Linguística
Published
2025-12-19
DOI
https://doi.org/10.25189/2675-4916.2025.v6.n4.id863
Primary Topic
Big Data and Digital Economy
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Plataforma da Diversidade Linguística Brasileira: dados linguísticos para uma IA brasileira

Kendra V. Dickinson, Amanda Post da Silveira, Brayna Conceição dos Santos Cardoso, Pedro Ricardo Bin et al.
Cadernos de Linguística
Big Data and Digital Economy
article

Plataforma da Diversidade Linguística Brasileira: dados linguísticos para uma IA brasileira

Kendra V. Dickinson, Amanda Post da Silveira, Brayna Conceição dos Santos Cardoso, Pedro Ricardo Bin, Marcus Garcia de Sene, Isabel de Oliveira e Silva Monguilhott, Andréia Silva Araújo, Iury Cleveston, Lilian Cristine Hübner, Sandro Marcío Drumond Alves Marengo, Valter de Carvalho Dias, Eliabe dos Santos Procópio, Hendrik Macedo, Ronice Müller de Quadros, Fabiane Cristina Altino, Marta Deysiane Alves Faria Sousa, Hadinei Ribeiro Batista, Raquel Meister Ko. Freitag, Marinete Rodrigues da Silva, Silvana Silva de Farias Araújo, Luma da Silva Miranda, Elisângela Nogueira Teixeira, Mailce Borges Mota, Túlio Sousa de Gois, André Britto de Carvalho, Miguel Patriarca Oliveira, Elisa Battisti, Alana de Santana Correia, Claudia Andrea Rosto Snichelotto, Monica Maria Guimarães Savedra, Marcia dos Santos Machado Vieira, Cleber Ataíde, Valéria Viana Sousa, Eduardo Cardoso Martins, Juliana Bertucci Barbosa
article en

Abstract

Generative artificial intelligence is based on large-scale language models (LLMs), which are trained with data most often collected without consent or in breach of copyright. LLMs are trained with billions of words and millions of parameters, but we don't know exactly which texts are selected in the training or which parameters are controlled. While unsupervised learning requires a large volume of data, demanding more and more computational costs and generating energy impacts, supervised learning with structured and tagged data can optimize this process; more than that: supervised learning with structured and tagged data resulting from language documentation projects can contribute directly to the National Artificial Intelligence Plan: “Develop advanced language models in Portuguese, with national data that encompasses our cultural, social and linguistic diversity, to strengthen sovereignty in AI.” In Brazil, in addition to Portuguese and its varieties, there are more than 250 other languages (indigenous, immigration, sign language), which are neglected in digital inclusion due to a lack of structured data. The consortium of laboratories and research groups in this INCT aims to prepare linguistic data for the training of LLMs, considering Brazil's linguistic diversity, with the development of a joint protocol for collecting linguistic data in the field, to be replicated in the groups and laboratories longitudinally, as well as transcription procedures, as well as procedures for transcribing, aligning and labeling linguistic data to create a data set that represents Brazilian linguistic diversity, and conducting studies on linguistic processing of diversity to fine-tune LLMs, helping to reduce asymmetries and prejudice resulting from training LLMs with translations from English.

Cadernos de LinguísticaVol. 6(4)
Eötvös Loránd University (HU), Universidade Federal do Rio de Janeiro (BR), Universidade Estadual de Londrina (BR), Universidade Federal do Rio Grande do Sul (BR), Instituto Federal da Bahia (BR), Universidade Federal do Acre (BR), Universidade Federal Fluminense (BR), Universidade Estadual de Campinas (UNICAMP) (BR), Universidade Federal de Sergipe (BR), Universidade Estadual de Feira de Santana (BR), Universidade Federal do Ceará (BR), Universidade Federal de Pernambuco (BR), Universidade Estadual de Santa Cruz (BR), Universidade Federal da Fronteira Sul (BR), Southwest Bahia State University (BR), Universidade Federal de Santa Catarina (BR), Universidade Federal de Jataí (BR), Rutgers Sexual and Reproductive Health and Rights (NL), Universidade Federal do Triângulo Mineiro (BR), Instituto Federal de Educação, Ciência e Tecnologia de Sergipe, Pontifícia Universidade Católica do Rio Grande do Sul (BR), Universidade Federal do Pará (BR), Universidade Federal do Amazonas (BR), Universidade de Pernambuco (BR), Universidade Federal de Alagoas (BR)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Conselho Nacional de Desenvolvimento Científico e Tecnológico
Quality Education
Openalex Percentile: Top 22%
Big Data and Digital Economy
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.