Key points are not available for this paper at this time.
Ethical issues surrounding artificial intelligence (AI) have raised wide concerns in today’s society. Addressing the issues in AI-generated texts can help mitigate the dilemmas that users encounter in AI applications. Previous reviews in this regard remain insufficient in terms of ethical risks in AI-generated texts, clear categorization of ethical issues in texts, and restricted opinions from a single disciplinary perspective. Accordingly, the current review is motivated to resolve these aspects by centering on the identification of AI-generation tools, the summarization of generated text types, and the classification of ethical issues on the basis of 57 studies. The results revealed 12 fields of application, 16 types of generated content, and six categories of investigative methods. Additionally, nine ethical problems and challenges were recognized, involving hallucination, reference and citation practices, copyright issues, academic misconduct, bias and discrimination, misinformation harm, instability, deprivation of self, and crisis of confidence.
Zhang et al. (Thu,) studied this question.