Abstract

The rapid spread of generative artificial intelligence(Generative AI) is reshaping how people seek and process information, increasing the need for AI literacy education. This study aims to identify the key elements of AI literacy education by examining the output generation structure of generative AI systems. To achieve this goal, the study theoretically analyzes how generative AI operates and produces responses, and explores how the structural characteristics of the generation process influence users’ information judgment. The analysis suggests that outputs produced by generative AI should not be understood merely as information delivery; rather, they represent discursive reconstructions influenced by probabilistic computations and platform-level design considerations. Drawing upon these structural insights, the study proposes several critical elements for AI literacy education within generative AI contexts: (1) input structure literacy, which involves recognizing the structure of prompts and problem framing; (2) computational process literacy, involving comprehension of computational procedures such as tokenization and embedding; (3) generation principle literacy, focused on grasping the probabilistic foundations underpinning text generation; (4) response structure literacy, which refers to understanding how responses are organized and presented; (5) platform structure literacy, emphasizing awareness of the mediating influence of platform-level design; and (6) critical evaluation literacy, highlighting the capacity to critically assess generated outputs. By systematically organizing these elements in relation to the output structure of generative AI, this study contributes to the conceptual framework of AI literacy education and suggests directions for AI literacy education in generative AI environments.