GenAI offers numerous applications in the context of research and can positively influence productivity.
GenAI can be used for various tasks in different phases of the research process, ranging from formulating a research question, conducting a literature review, writing a project proposal, data collection and analysis, reporting results, to reviewing project proposals and manuscripts.
The use of GenAI carries relatively less risk for certain tasks, while it can be more risky for others. The general rule is: the more responsibility you place on GenAI, the more verification, control, and accountability are needed for its output.
Discover what researchers around the world think about the use of AI in their work in this report from 2024 published by Elsevier.
Below are several examples of how GenAI can be utilised in research. Always adhere to the guidelines for responsible use (see the relevant tab in this library guide)! Here, we refer to GenAI, AI tools, AI chatbots, or generative language models in general. Under the Tools tab in this library guide, you will find an overview of the specific tools that can be employed. We only refer to specific tools below if they add value (KU Leuven, 2023a).
The use of GenAI can be helpful when conducting a literature review, for example, because it is time-saving. However, good literature review does not result in a simple enumeration or summary of all available literature, but in a critical discussion of the most relevant information. Therefore, AI, given its limitations (see the tab "What is it?" in this library guide), is not suitable for fully conducting a literature review! But when you adhere to the guidelines for responsible use (see the tab "Responsible Use" in this library guide), AI can support discovering relevant literature, extracting key insights, structuring the content of the literature and your own research, resulting in a more streamlined and effective work process (The Education University of Hong Kong, n.d.).
Below, we discuss the use of GenAI in literature searching (Boetje, 2023; Zuyd Library, 2023). We base this on these steps. For each step, advice is given on how to apply GenAI and what the pitfalls are. Table of contents: 1. Research question, 2. Search strategy, 3. Information sources, 4. Evaluating information, 5. Processing information. Below, we refer to GenAI, AI tools, AI chatbots, or generative language models in general. Under the Tools tab in this library guide, you will find an overview of the specific tools that can be utilised. We only refer to specific tools below if they add value. Additionally, through this link (Dutch), you will find a list of AI tools applicable for conducting literature research, compiled by the Information Literacy working group of the collaborating university and college libraries in the Netherlands.
1. Research question
During this step, you explore the topic, then narrow it down to what interests you, and finally formulate a research question (for example, according to the SMART model: a SMART research question is Specific, Measurable, Achievable, Relevant, and Time-bound).
► GenAI can be useful for exploring specific topics, for instance, through automatically generated mind maps using edrawmind.com. Alternatively, you can ask an AI chatbot for (general) information on a particular subject.
► GenAI can assist in clarifying the assignment description and the requirements your literature research must meet by engaging in conversation with an AI chatbot and asking questions based on the assignment description from your teacher or client.
► AI-generated suggestions for relevant (research) questions around a theme can help in finding a solution or answer.
► Explore which concepts and theories are relevant to your research. An AI chatbot can clarify these for you.
► You might ask an AI chatbot to formulate a research question according to the SMART criteria.
► GenAI cannot help you determine your information needs. However, you can use the tool to gain inspiration when formulating your research question or to broadly explore your topic
2. Search strategy
During this step, you search for as many relevant search terms as possible that align with your research question and determine which search methods you will apply.
► AI can suggest alternative search terms that might be more relevant to your query. Specific tools even provide entire mind maps of related keywords around a theme (for example, ContextMinds).
► Litmaps can be a useful tool for applying the citation and snowball method. This tool makes it clear which articles are connected through citations.
► Using AI to determine search terms does not replace the need to consult additional sources, such as a (synonym) dictionary or thesaurus, or to check the jargon used by authors or experts in your field.
► Be aware that GenAI is not trained to formulate search queries for database searches. You will need to edit and rephrase the output it provides.
3. Information sources
In this step, you identify the databases and sources you want to search.
► You can ask an AI chatbot for clarification about a specific database or source.
► It is advisable to use AI only during the exploratory and processing phases and conduct the literature research itself through a scientific search engine or in a subject-specific database that matches your research question. AI cannot provide a complete or exhaustive overview of available subject-specific databases or sources. Moreover, the tool cannot tell you whether you have access to the information in a database. The full overview of databases and sources (with access information) can be found on Greeni.
► AI cannot provide information from after the date their training was completed. In contrast, the content of most databases is updated daily with new information.
4. Evaluating information
During this step, you evaluate the search results you have found.
► AI can speed up the selection and scanning of information by generating automatic summaries of PDFs, texts, and videos.
► Plugins like ExplainAI can help explain unfamiliar words in web texts using AI.
► AI tools can assist in automatically generating a mind map or context map from one or more sources, making the relationships between different ideas and themes visible (e.g., ContextMinds).
► ChatGPT cannot assess or select which information is relevant for you. You will need to do this yourself.
► ChatGPT also lacks insight into the context (such as the level) in which the literature research is conducted. Therefore, as a researcher or student, you can only decide whether information is of sufficient quality to use in your work.
5. Processing information
In this step, you will process the information you have gathered into your final product. You will ensure that your final product is verifiable and reproducible. This includes providing correct source citations.
► You can use AI to clarify difficult texts, as it is good at rephrasing text.
► AI can create automatic transcriptions of audio clips or convert text into other forms such as videos or images.
► AI can assist in creating a coherent narrative from information gathered from various sources.
► AI can help organize collected information in a logical order.
► AI can provide a framework for what the final product should look like, for example, by asking for a sample structure.
► AI can assist in creating charts or designing the final product.
► AI can help improve the final product by providing constructive feedback and assisting with text revisions, thereby enhancing the quality of the final product.
► AI cannot or will not provide sources for the texts it writes. Furthermore, something strange happens when you ask AI to create a bibliography on a specific topic. It will come up with a list of sources that initially seem accurate. However, upon closer inspection, these sources are often fake and completely fabricated. Therefore, compile your source list yourself.
► ChatGPT also cannot create reference lists according to specific guidelines. The reference program Mendeley is a much better choice for compiling a source list in a specific reference style. The preferred style at Green UAS is APA. Use the guide The APA guidelines explained.