<![CDATA[Generative artificial intelligence (AI) can produce persuasive copywriting while adding claims or details that are not supported by the source text, creating a need for critical evaluation of AI-generated information in education. This study aims to identify information changes in human-authored and AI-generated copywriting and to formulate pedagogical implications for critical AI literacy in elementary education. A qualitative document analysis compared ten human-authored source texts with seventy outputs generated by seven AI models in an Indonesian small-business marketing corpus. The analysis examined information retained, omitted, transformed, or added, particularly claims about product attributes, health benefits, prices, stock availability, and promotional conditions. Selected examples show that some AI outputs introduced details such as pesticide-free status, suitability for complementary feeding, therapeutic benefits, discounts, stock limits, and ordering schedules that were not established in the source texts. These additions cannot automatically be classified as false; rather, they require independent verification before being treated as factual. The findings suggest that carefully selected copywriting examples can support age-appropriate activities in which pupils distinguish source information from AI additions, examine persuasive language, and justify claims using evidence. The study does not measure student learning outcomes, and its pedagogical implications therefore require classroom testing. Before instructional use, the corpus should be fully audited, including verification of a reconstructed source record and review of excluded residual rows.]]>
Copyrights © 2025
