Abstract

The purpose of this study was to investigate whether humans can distinguish between human-generated and artificial intelligence (AI) chatbot-generated text, and to identify key differences between the two. The dependent variable was the participants’ accuracy in identifying each text, while the independent variable was the group of participants. Controlled variables included the type of passages and the time allocated per participant. The hypothesis proposed that participants would not be able to accurately differentiate between AI- and human-generated text.

A descriptive and mixed-method design was employed. A total of 456 participants completed questionnaires containing 10 passages. Each participant was required to identify whether each passage was human- or AI-generated within 60 seconds. At the end of the questionnaire, participants provided a brief explanation of the factors that influenced their decisions. Data were recorded in tables and analysed using bar graphs and pie charts.

Results indicated that 99.11% of participants were unable to accurately distinguish between AI- and human-generated text. Only 4 participants correctly identified all passages, all of whom attended schools in urban areas. These findings reject the initial hypothesis, as not all participants failed to identify AI-generated text. Nevertheless, the overall ability to differentiate between the two remained low. The data further suggested a statistically significant difference in recognition ability between learners from urban and rural schools, implying that exposure to digital technologies may enhance the capacity to detect AI-generated content.