Idea development and research design

Use of generative AI systems during the process of idea development and research design

There are various ways to find research topics, formulate research questions, and define appropriate methods, including consulting existing literature, brainstorming with colleagues and supervisors, and attending academic conferences. While these traditional sources remain valuable, generative AI systems can complement them and even serve as an excellent starting point for research.

Nonetheless, before using generative AI systems to support idea development and research design, it is essential to clearly define your area of interest and topic. Furthermore, the researcher must have a basic understanding of the subject to critically evaluate the AI system’s outputs. This preparation increases the effectiveness of interactions with these systems, as the researcher should be able to verify the information and, in particular, recognize citations of non-existent research works.

!
Read below to understand how generative AI systems can support researchers in idea development and research design.



!     To complete this task, follow these steps:
(1) Define an area of interest and a topic you want to work on.
(2) Then insert it into the following prompt: “Create 10 possible research questions on [Topic X] and rank them according to their relevance for [Research Field Y].”
(3) Enter the prompt into a generative AI system of your choice, for example ChatGPT, Gemini, Claude, Perplexity, Llama, DeepSeek. However, we recommend using the FU-internal AI chatbot if you have concerns about data protection and IT security. Read more here about responsible use of LLMs.
(4) Evaluate the response provided by the AI system and select one of the proposed research questions.
(5) Then enter a new prompt: “Suggest possible research methods to answer the following [Research Question X], and rank them according to rigor, quality, and reliability of evidence.”

 
 
Note: This is just one way in which generative AI systems can support researchers during the process of idea generation and research design. Researching other approaches for this task is recommended.
 
Possible Risks
Researchers must be aware of the significant risks associated with using AI systems:
 
  • Generative AI systems can invent information, so-called "hallucinations" (e.g., non-existent research papers).
  • Users may be tempted to copy and paste generated research drafts directly, which could reduce the learning effect for early-career researchers that comes from experimentation and trial-and-error.
 
Risk Mitigation
  • Be a critical user of AI systems and evaluate their outputs carefully.
  • Consult more experienced researchers to discuss the reliability and feasibility of suggestions provided by AI systems.
  • Use the generated outputs as a supplementary/supporting tool, not as a replacement for the process of idea generation and research design.