Automated Audio Description in the Teaching of Spanish as a Foreign Language: Challenges and Perspectives

desafíos y perspectivas

Abstract

Since 2020, the widespread adoption of artificial intelligence (AI) tools, understood as technologies capable of simulating human abilities, has intensified. In this context, discussions about their impact across different fields, particularly in education, have expanded. This article aims to analyze automated audio descriptions generated by AI tools from images in the Spanish language textbook Gente Única, considering criteria related to character description, setting, contextualization, and pedagogical appropriateness. The study examined audio descriptions produced by the platforms ChatGPT, Zapia, Be My Eyes, Lookout, and Copilot, based on the criteria proposed by Motta (2016) and Naves et al. (2016). This is a qualitative, exploratory study conducted through a case study approach (Chizzotti, 2006). The findings reveal limitations of AI systems regarding the pedagogical, cultural, and linguistic sensitivity required for audio description in educational materials, reinforcing the importance of professional audio describers in ensuring appropriate accessibility within educational contexts.