In this study, we examined online news comment activities that are used as a key means of public opinion and interaction in modern society. We extracted keywords in comments using big data and sorted them according to frequency. Through this process, a chatbot was designed to check the ranking and contents of comments based on keywords represent the most important issues in the comment population and search for classified comments according to keywords before users read the news. Through experiments using this chatbot, we compared and analyzed the nature of the comments in the existing comment system and the characteristics of the system experience. Results shows that the chatbot designed in this study can provide more information than existing comment presentation systems, search for comments by type, and check more information than articles. In addition, assessing the nature of the existing comments reduced the number of sensational, non-slang-oriented comments.