THE AUDIO DESCRIPTION PRODUCED BY ARTIFICIAL INTELLIGENCE TOOLS: A STUDY BASED ON IMAGES FROM A SPANISH TEXTBOOK AS A FOREIGN LANGUAGE
Audiodescription. Artificial Intelligence. Spanish as a Foreign Language. Didactic Material.
Artificial intelligence has sought to position itself as a support tool for promoting accessibility for people with visual impairments. However, the lack of adequate audio description in educational materials still represents a significant challenge, limiting blind students' access to essential visual content in the learning process—especially in fields such as language teaching, where vocabulary and visual context play a fundamental role in knowledge construction. In this context, the general objective of this research is to evaluate image audio descriptions from the Spanish textbook Gente Única generated by artificial intelligence tools, based on audio description guidelines (Motta, 2016). The theoretical framework approaches audio description as a mode of intersemiotic translation, essential for the inclusion of people with visual impairments in language learning. The study also discusses the role of AI as a support tool for accessibility, highlighting its potential to produce automatic audio descriptions. A qualitative, exploratory-descriptive methodological approach is adopted, through which audio descriptions generated by AI tools—such as ChatGPT, Zapia, Be My Eyes, Lookout, and Copilot—are evaluated based on criteria such as the description of characters, settings, use of verb tense, and interpretative impartiality. The results indicate that although AIs show potential to generate audio descriptions, there are significant limitations, such as a lack of detail in describing characters and settings, the absence of frame identification, and the omission of contextual information.