최신자료
Executive Summary: The Performance of Artificial Intelligence in the Use of Indigenous American Languages
This publication analyzes the performance of artificial intelligence (AI) models when interacting in Indigenous American languages. The study evaluates seven representative languages from Latin America using five different language models, identifying a significant performance gap compared to major languages such as Spanish or Catalan. Through methodologies like the Multidimensional Quality Metrics (MQM) and the Multi-Task Language Understanding (MMLU), the report measures the linguistic, executive, and behavioral performance of the models, revealing significant limitations in comprehension, expression, and cultural adaptation. Additionally, the report examines the scarcity of digital data and linguistic tools available in Indigenous languages, which limits the training of AI models. It presents 21 strategies to promote technological inclusion, ranging from the creation of international consortia to the promotion of hackathons and data collection initiatives. The document concludes with a clear action plan to reduce the technological gap and promote the fair and responsible use of AI in multilingual contexts. * A more detailed version of this publication is available at: http://dx.doi.org/10.18235/0013542
많이 본 자료
![[2025.07] 대출규제](/userdata/main/incContent/sideLink/23/aaaqQGcnBTcOc8zW9uLNz20251015165742zqfnw.png)
![[사람이 되고 싶어요2] 환율의 변화](/userdata/main/incContent/sideLink/22/aaaqQGcnBTcOc8zW9uLNz20251015165408vtzoa.jpg)
