<![CDATA[This study examines the variation in children’s spoken language, with a particular focus on the use of swear words. It compares how swear words are defined and analyzed by Artificial Intelligence (Perplexity AI) and the electronic form of Kamus Besar Bahasa Indonesia (KBBI). The research adopts a sociolinguistic framework, drawing on Steven Pinker's (2005) theories of swearing, Anderson and Trudgill’s (1992) work on language variation, and Natural Language Processing (NLP) techniques. By integrating these approaches, the study aims to explore the variation and transformation of swear words among children. Data were collected through primary sources, including direct observations and voice recordings of children’s conversations. The analysis involved categorizing the swear words, identifying their functions, analyzing their use in context, and comparing their definitions in Perplexity AI and KBBI. The study found that the online KBBI struggles to keep up with the rapid development of swear words used by children, whereas Perplexity AI offers a more adaptive and accurate representation of these terms. Seventeen swear words were identified and categorized into four groups: humorous (6 words), expletive (5 words), auxiliary (3 words), and abusive swearing (3 words). These words performed 13 distinct functions in children’s conversations. The findings highlight the limitations of traditional dictionaries in capturing contemporary language, while also emphasizing the potential of AI tools in documenting and analyzing evolving linguistic trends. This study also finds the transformation of swear words based on cultural and social context.]]>
