My Experience with ChatGPT for Research

As a researcher in social science, I am always on the lookout for new tools and techniques to improve my research process. Recently, I have been experimenting with ChatGPT, the generative AI developed by OpenAI, and it has significantly transformed the way I generate ideas, synthesize information, and draft content. In this article, I will be sharing my personal experience of using ChatGPT for various stages of my business research projects and the remarkable impact it has had on my work.

Generating Research Ideas with ChatGPT

One of the most challenging aspects of any research project is identifying original research ideas. From a philosophical perspective, this process of ideation often involves combining and transforming ideas we already know. ChatGPT has proven usable in this regard. By providing a well-structured prompt that includes the topic, research question, and key aspects to consider, ChatGPT generates a range of original ideas that serve as a starting point for my research.

A prompt template I developed recently was:

Write a formal and technical introduction to a research paper on [insert topics] in economics and business. Adopt the tone and style of an esteemed academic researcher and address the research question of [insert research question]. 

The ideas should be innovative and based on current literature or data. Consider the following aspects while generating ideas:
- [insert one aspect for ChatGPT to consider]
- [insert another aspect for ChatGPT to consider]

You may use your own methods and design and do not need to follow the talking points word by word. Feel free to make up details including numbers, methods, and design. Don't worry about the correctness and feasibility of your statements, as I will correct them manually if necessary.

This prompt helped me organize my thoughts and explore new perspectives on the research topics I am interested in. However, it is essential to provide more guidance to ChatGPT; otherwise, it is likely to only spit out generic and mediocre research proposals.

Integrating ChatGPT into My Research Workflow

Beyond generating research ideas, I have found ChatGPT to be a versatile tool that has streamlined my research process in numerous ways:

  1. Literature Reviews: many emerging tools powered by LLMs, such as SciSpace, have expedited my literature review process by summarizing relevant articles and identifying key themes and trends. This has saved me a significant amount of time and allowed me to better identify research gaps and areas that need further investigation.
  2. Hypotheses and Methodologies: Instead of relying on ChatGPT to design the methodologies or research design, I have found it to be an invaluable tool for receiving constructive feedback on my research plans. By asking ChatGPT to act as a peer reviewer and provide questions or suggestions on my proposed research design, I have been able to formulate appropriate plans for my projects.
  3. Drafting and Editing: As a language model, ChatGPT is a killer app for copywriting sections of my research papers and editing them for clarity, coherence, and readability. This has not only improved the quality of my writing, but also significantly reduced the time spent on revisions.


Incorporating ChatGPT into my business research workflow has had a profound impact on my productivity and the quality of my work. Its versatility and ease of use have allowed me to streamline my research process and focus on the critical aspects of analysis and interpretation.

That being said, current AI still suffers from the hallucination issue (Wikipedia), and we should always take the output from ChatGPT with a pinch of salt. We should view it as a writing assistant, rather than a ghostwriter who can do everything for us. This post made by Terence Tao just about summarizes my experience with ChatGPT:

I’m beginning to think that AI tools obey a variant of Cunningham’s law: “The best way to get the right answer on the Internet is not to ask a question; it’s to post the wrong answer”. There have been multiple times now where I have asked GPT for some text for some task, was dissatisfied with the output, but then was spurred by that dissatisfaction to write some text more to my liking, perhaps just borrowing a sentence or two from the GPT-produced text. Without that initial text serving as a prompt, though, I don’t think I would have been as inspired or motivated to generate my own version.

I highly recommend fellow researchers to explore the potential of ChatGPT and embrace the power of AI in their research endeavors. By harnessing its capabilities, we can unlock new opportunities and increase efficiency in our work.