ChatGPT in Academic Writing and Publishing: A Comprehensive Guide

ChatGPT in Research Conception (I)

2.1. Research Conception

Research concept refers to a researcher’s understanding of what research is and what research is. Higher education researchers may have different concepts of research practices [7]. From research questions to research design, from literature review to meta-analysis, ChatGPT and other conversation AI chatbots can assist in all stages of research conception and methodology (Figure 2.).

Figure 2: ChatGPT's functions in research conception.

ChatGPT can help researchers generate research question ideas by providing suggestions based on user input. For example, when researchers enter a topic such as "The use of nanomaterials for green energy production", ChatGPT generates a list of potential research questions related to that topic. This feature is particularly useful for researchers who are new to the field or who struggle to develop research questions.

2.1.1. Hypothesis generation

ChatGPT and similar language models are useful for various scientific research applications. ChatGPT can identify and propose hypotheses for further research by analysing a large amount of scientific literature. Additionally, it can help researchers brainstorm broader research questions and identify knowledge gaps in their fields. ChatGPT is a potential use for the generation of hypotheses for research articles to generate possible research questions and hypotheses based on existing literature and data [8, 9]. For example, authors and researchers can ask ChatGPT to suggest a research question or hypothesis related to a topic and then evaluate and improve it using their knowledge and expertise. The use of  ChatGPT helps researchers explore new hypotheses by creating plausible explanation conversations [10].

Another possible application of ChatGPT to generate hypotheses for research articles is to use it as a feedback tool to test and improve your hypotheses [11]. For example, ChatGPT can test your hypothesis, provide counterarguments or alternative explanations, or propose ways to test or falsify your hypothesis. ChatGPT helps to strengthen hypotheses, identify potential errors and limitations, and improve the validity and reliability of research [12].

ChatGPT can help researchers explore new ideas by suggesting connections between seemingly incompatible concepts. Researchers can engage in an open dialogue with the artificial intelligence system to explore new concepts and theories in their fields of interest (Figure 3). AI models can also help identify assumptions, arguments, gaps, or other shortcomings in scientific literature.

Figure 3. ChatGPT response to: Gives hypotheses and idea exploration about using chitosan in wound healing. 

2.1.2. Literature review

The comprehensive review of the literature is a key part of academic and scientific research. Conducting a literature review is an important step in the research process, but it is time consuming and difficult [13]. With the emergence of artificial intelligence (AI), researchers have access to tools that can help them more efficiently review literature. Chatbots or conversational agents are a promising tool for this purpose.

ChatGPT can help researchers automate the search for relevant publications, extract key information, and synthesise results from multiple studies [11]. ChatGPT provides a summary of current research on this subject and an overview of any literature gaps. Researchers can also ask ChatGPT to filter the literature by criteria such as date, author, journal, or keywords. This capability can save researchers a lot of time and effort and allow them to focus on interpreting and applying research results.

Another advantage of ChatGPT is its ability to help review the literature (Figure 4). ChatGPT helps researchers identify relevant literature by analysing user input and creating a list of related articles. ChatGPT analyses the inputs and generates a list of articles related to the research questions and topics [14]. This feature can save researchers a significant amount of time compared to manually searching for literature. ChatGPT can also help evaluate the quality of articles by analysing abstracts and identifying key findings.

By analysing articles and literature, ChatGPT can identify areas in which gaps in a specific topic or additional research or further research is needed [15]. This function can be especially useful for researchers who are new to the field or looking for gaps in research in a particular field.

Figure 4. ChatGPT response to: Conduct literature review for using nanomaterials in architecture.

2.1.3. Meta-analysis

Meta-analysis is a powerful tool for synthesising data from several research projects to better understand specific research issues [16]. One of the greatest challenges in meta-analysis is the sheer volume of data to be extracted and analysed [17]. With the emergence of artificial intelligence (AI), researchers have access to tools that help them perform meta-analysis more efficiently. Chatbots, or conversational agents, have become increasingly popular as a tool for meta-analysis in recent years.

 Researchers and authors can use OpenAI's large-scale language model ChatGPT for meta-analysis of research papers (Figure 5). One of the main benefits of using ChatGPT to perform meta-analysis is the ability to help in the data extraction process. ChatGPT analyses articles and extracts relevant data, such as effect size, sample size, and study design. The use of AI model in data extraction saves researchers a lot of time compared to manually extracting data. In addition, ChatGPT can help identify relevant articles by analysing the researcher's input and generating a list of articles relevant to the research topic.

Another advantage of using ChatGPT in meta-analysis is its ability to help with statistical analysis. ChatGPT can perform statistical tests [15] such as calculating effect sizes and conducting heterogeneous analyses, thereby saving researchers considerable time compared to manual analyses. ChatGPT can also help researchers identify potential publications' biases by analysing articles and identifying any possible sources of biases, such as selective reporting and publication. This ChatGPT's function is particularly useful for researchers conducting meta-analysis on controversial subjects or in areas where published biases are common.

Figure 5. ChatGPT response to: Conduct Meta-analysis for using mRNA in COVID-19 vaccine.