Ethics and biases in the use of ChatGPT for academic research

Abstract

This study investigates the ethical considerations and biases in the use of ChatGPT for academic research, focusing on its acceptance, perceived usefulness and the impact on intentions for future use among university students. A quantitative approach was used, collecting data from 5,000 participants at 20 universities. The relationships between key variables were analysed using structural equation modelling (SEM) with SmartPLS, and reliability and validity were assessed using Cronbach's alpha and AVE. The results revealed that ease of use (coefficient = 0.389, p < 0.05) and content reliability (coefficient = 0.530, p < 0.01) have positive and significant effects on the intention for future use of ChatGPT. However, ethical considerations (coefficient = -0.047, p > 0.05) and perceived impartiality (coefficient = 0.120, p > 0.05) did not show significant influence. Furthermore, the acceptance of ChatGPT significantly impacts academic research (coefficient = 0.936, p < 0.001). It is concluded that ChatGPT offers significant benefits for academic research, such as improved efficiency and productivity, but its adoption must be accompanied by ethical guidelines to mitigate risks related to authorship and originality. Regarding practical implications, higher education institutions should establish clear policies and training programmes to promote the responsible and ethical use of artificial intelligence tools such as ChatGPT. These measures will ensure that students can take advantage of these technologies effectively without compromising academic integrity.