Potential Drawbacks of Automated Paraphrasing Tools: Performance of QuillBot in Paraphrasing of Headline Discourse

Abstract

Revolution in development of automated paraphrasing tools (APTs) has brought drastic change in paraphrasing text particularly through QuillBot. Studies indicate that performance of this tool is highly satisfactory. However, QuillBot has certain drawbacks that affect the paraphrased text such as it neutralizes the text by replacing subject with general statement, breaks appealing potential of text, confuses the context of the headline text, and increases number of words of the paraphrased text. Thus, the objective of this study was to investigate the performance of QuillBot. So, headlines from The Washington Post and The Times were selected randomly as data and were analysed on self-designed conceptual framework based on four standards namely neutralization, de-contextualization, semantic difference and violation of brevity of the text. The study found that there is 50% accuracy of QuillBot with respect to standard 1 and standard 2. Standard 3 showed 40% accuracy while standard 4 showed 99% accuracy.