Artificial Intelligence (AI) has experienced exponential development in recent years. ChatGPT developed by OpenAI has succeeded in providing an important benchmark in language development models that are able to understand how humans think. With users reaching 100 million worldwide, chatGPT is considered capable of answering society's current needs for fast and accurate information. It is interesting to examine how linguistic variations include input from users and the feedback given. This research uses a case study method by entering linguistic variations as text input and looking at linguistic aspects of GPT chat responses. The research results show that overall, chatGPT is very dependent on user text input in providing feedback, the more complete the commands and information entered, the more specific the response given by chatGPT. Another finding was that chatGPT was able to produce responses with diction choices with academic and informative nuances, but on the other hand there was minimal variation in synonyms; and be able to understand and apply Indonesian language rules well and correctly in the resulting text
Copyrights © 2023