Benefits and challenges of using generative artificial intelligence tools in scientific research: A systematic review

Abstract

Purpose: The present study aimed to investigate the benefits and challenges of using generative artificial intelligence tools in scientific research and production using a systematic review method.

Method: The present study was conducted using a systematic review method. For this purpose, foreign scientific databases including Google Scholar, Science Direct, Web of Science, and Scopus were reviewed. After searching for articles and applying screening based on the inclusion and exclusion criteria for the study, including publication of articles in English, availability of full and downloadable text of articles, publication of articles in reputable scientific journals, and publication of articles between 2000 and 2024, a total of 41 articles related to the topic were selected for qualitative meta-analysis. The reason for selecting foreign scientific databases was the novelty of the topic and the lack of domestic research on this topic.

In order to conduct a qualitative meta-analysis, first, the texts related to 41 related articles were carefully studied. Then, open and axial codes were extracted. In the open coding stage, all categories were identified and in the axial coding neighborhood, the categories were placed in separate categories based on their similarities and differences. The goal of this stage was to establish connections between the categories identified in the open coding stage and it is called axial coding because the coding is done around a category. In this stage, the opportunities of using generative artificial intelligence tools in research and the challenges of using generative artificial intelligence tools in research were identified as two axial codes and an attempt was made to place other open codes around these two axial codes. In total, one main category, two subcategories (axial codes) and 18 subcategories were identified.

In order to conduct this review, all studies conducted on the subject using relevant keywords were extracted from the Scopus, Google Scholar, Science Direct, and Web of Science databases, and after screening, 41 articles related to the subject were reviewed.

Findings: The research findings can be presented in the form of descriptive and analytical findings. The descriptive findings of this study showed that most of the articles published on the subject of the present study were written in the last two years (2023 and 2024) and simultaneously with the development of the use of generative artificial intelligence tools in scientific research, which indicates the emergence and novelty of the subject under study. The method used in most of the articles reviewed is the review and documentary method, and the ratio of field research (quantitative, qualitative, and mixed) to documentary research (review and systematic review) is lower in this area. The geographical distribution of the authors responsible for the articles studied showed that the largest number of articles were from the Americas, Europe, Asia, Australia, and Africa, respectively. The scope of the articles studied was, in order of abundance, in the areas of opportunities and challenges arising from the use of generative artificial intelligence tools in research and education, the use of generative artificial intelligence in research, the use of generative artificial intelligence in education, the use of generative artificial intelligence in research and education (combined use), and the social, legal, and ethical implications arising from the use of generative artificial intelligence. The analytical findings of the study showed that the use of generative artificial intelligence tools in the field of scientific research has created numerous opportunities and challenges. Among the opportunities for using these tools are increasing justice in scientific writing, increasing the speed and accuracy in collecting, organizing, and analyzing large volumes of scientific data, facilitating and comprehensive digital access, improving academic services, improving the accuracy and quality of scientific writing, accelerating scientific refereeing processes, making scientific research more attractive and understandable, and creating new forms of scientific writing and research. The findings also showed that the challenges arising from the use of the aforementioned tools in scientific production include: algorithmic bias and discrimination, issues related to transparency and privacy, production of false information, lack of originality, unreliability, lack of accountability and fear of plagiarism, increasing the rate of production of poor-quality, incomplete or irrelevant articles, legal concerns regarding intellectual property rights and copyright, fear of misinterpretation and misunderstanding by users, and the loss or replacement of many jobs in the field of research.

Conclusion: The subject under study is new and not much research has been conducted using field methods in this regard. In addition, the results of the present study showed that the use of generative artificial intelligence technology in scientific research has numerous advantages and challenges. Strengthening the advantages and reducing the weaknesses and challenges of using the aforementioned tools requires policy-making in line with the responsible use of this technology in coordination and cooperation with the stakeholders involved and the various actors in this ecosystem.