Evaluation of an AI translator for adapting curriculum content to indigenous languages in primary schools

Surgey Bolivia Caicedo
Additional contact information
Surgey Bolivia Caicedo: Universidad de Pamplona: Norte de Santander

Applied Journal of Digital Innovation, 2025, vol. 1, 4

Abstract: This study critically evaluates the potential of Artificial Intelligence (AI) translators for adapting curricular content to indigenous languages in primary education. Using a mixed-methods approach (quantitative and qualitative) that included technical evaluation of AI models and perspectives from specialists, teachers, and indigenous communities, the findings reveal a fundamental paradox. Although specialized AI systems show technical improvements in fluency and syntactic accuracy, they exhibit a structural insufficiency in cultural-pedagogical relevance, the core of meaningful adaptation. The most serious errors were related to cultural correspondence and linguistic calques. The research identifies risks such as "hidden standardization" (amplification of majority dialectal variants) and underscores that ethical governance and community sovereignty over data and processes are non-negotiable prerequisites for any implementation. It is concluded that AI cannot replace the creative and situated role of the human translator, but could function as a technical assistant within a "Community-Supervised Hybrid Model", where technology is subordinate to indigenous agency and control. Feasibility depends less on technical advancement and more on frameworks of epistemic justice and digital decoloniality.

Keywords: Machine translation; indigenous languages; primary education; cultural relevance; linguistic justice; ethical. (search for similar items in EconPapers)
Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://aj.reincisol.com/index.php/aj/article/view/4 Abstract page (text/html)
https://aj.reincisol.com/index.php/aj/article/download/4/14 Full text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ajd:ajdinn:v:1:y:2025:id:4

DOI: 10.59282/aj.2025.1.4

Access Statistics for this article

More articles in Applied Journal of Digital Innovation from Reincisol Publishing
Bibliographic data for series maintained by Homero Lastra ().