An Initial Investigation on Originality of Text Generated by Generative AIs Like ChatGPT

Generative Artificial Intelligence (GAI) is an artificial intelligence that is capable of generating new and original content. With ChatGPT going public in late 2022, GAI is receiving an immense attention. While people are getting very excited about impressive capabilities of this technology in creating textual content such as letters, reports and essays, there are concerns about the originality of the content produced by GAI, especially in the higher education sector.
In response to this concern, this study takes an initial step through examining the originality reports of 60 essays generated by 5 different GAIs, namely, ChatGPT, YouChat, Chatosonic AI, Perplexity AI, and ChatGenie. To collect data, the study first chose one topic from each of four subject areas, namely, social science, mental health, organizational studies and computer science. Next, each of the 5 above-mentioned GAIs were asked to generate essays of 200, 500 and 1000 words for each topic, resulting in a total of 60 essays which were then submitted to the Turnitin plagiarism detection software. The results showed that while 16 essays were found to be plagiarism free, the other 44 had certain amount of plagiarism ranging from 3% to 47%. A correlation was observed between the word count of the text generated by GAI and the plagiarism percentage, with lower word counts showing a higher plagiarism percentage. Limitations of the study as well as notes for practitioners are reported as well.