This activity introduces mind mapping, a practice in which students explore and develop their research topic and determine its core concepts. As students explore, the instructor provides instruction to help students identify questions by type and to search thoughtfully and strategically for sources from contextual background information to academic scholarship. This activity may be used to introduce BU Libraries Search and search string formation. This workshop is most impactful when delivered as students begin a research project. 

Note: This was designed as a 90 minute workshop, but can be condensed as needed. Faculty should feel free to display some or all of the slides created to accompany this workshop with students.

Before beginning, determine whether this will be an individual or team activity, and decide what materials you will need:

  • Individual: Students will spend the whole time working on their individual mind maps related to their research assignment. Print mind map templates for each student.
  • Team: Students create a mindmap for a shared topic (e.g., urban green space). Switch to individual work when they write their research question on the “refine your topic” slide. Provide one giant sticky note (25 x 30”) for each team.

Objective

Practice exploratory research and topic development; develop and use keywords to search for information; Reflect on ethical AI usage as a thought partner when doing research.

Key Terms

AI research; information literacy; mind mapping.

Part 1: Explore Your Topic

  1. Brainstorm 1-3 question(s) that you have about the topic. Quality does not matter here, questions don’t need to be complex.
  2. Replicate the brainstorming using an AI tool of choice – generate 3-5 more questions and write them down in the questions box on the mind map.
  3. Explain types of questions (background and research) and how they are used in the research process.
  4. Discuss human-generated vs. AI-generated questions. Connect to question types already discussed, and notice trends in AI-generated questions.

Part 2: Mind Mapping

  1. Using the questions to guide the mapping, write down ideas to explore on the map, then draw lines to connect them to the main topic.
  2. Encourage students to look for background information while mapping. Utilize online and library sources. If they use AI for background information, be sure to confirm with another reliable source.
  3. Share the mind maps

    Part 3: Refine Your Topic

    1. Take this to the next level: The same way we used questions to guide our map, now write a statement that summarizes your research topic / research question. This will be used to guide your search for information
      • I’m researching ___ because I want to find out why/what/how ____  (3 min)
    2. Share with a neighbor. Emphasize that a research question will change throughout the research process so it doesn’t have to be perfect. 
    3. Generate keywords: Explain why keywords are important and compare it to prompt engineering for GenAI. Take 2 minutes to identify the core concepts in your research statement, and any related concepts that come to mind. (Hint: these core concepts are more than likely already on the mind map)
    4. Use an AI tool of choice with the prompt: What are the 5 most important terms used by researchers when examining [insert research topic]? 
    5. Discuss and compare the human-generated search terms with the AI-generated ones. Which results feel more relevant to your question? What would you do next to improve your results? How would you change your prompt?

    Optional: Library Expansion Activity

    1. Demonstrate the BU Libraries Search and have students try out the search terms in a library database to see what they find:
      • AI tools accept natural language prompts like Google and other search interfaces you’re familiar with
        • The results are generated through predictions about what will be most useful to answer the prompt based on the information that the tool has been fed and trained on, so what you see is influenced by what others have said about a topic, the data that the LLM has been fed, and the algorithm itself. 
        • Library tools don’t accept natural language prompts, they use a technique called keyword searching which means that in order to search effectively, you need to break your topic down into keywords before you even think about searching
        • Something else to know is that anybody using the same keyword combination will get the same results in the same order based on relevance. 
      • Explain how databases think – need to use keywords linked together with Boolean Operators. 
    2. Think about the ordering of search results. In a Library database, the results that you get are ordered by relevance as determined by the frequency of keywords in parts of the item record. Example with whatever is the top result on my search. The goal is discovery and browsing, replicating the experience that you might get in a physical library space. Google ranks search results based on other user behavior and also who is paying them to rank results higher. Yes, term frequency plays a role but not in the same way that a library database does – there’s a lot of other things that come into play like your geographic location, your own previous search behavior, etc. Large Language models are able to make somewhat accurate predictions about what is relevant based on data that it is fed, so you don’t experience discovery, you get a recommendation based on what other people have used or produced. The information that is prioritized will be the information that is most prevalent meaning that the voices of the old white man scholarly canon – the loudest voices – are the kind that your results will reflect.

    Part 4: Reflect

    1. Think-pair-share about the benefits and hazards of using an AI-assisted research process.
      • How did this process feel? What did you learn from it?
    2. Food for thought: Ask yourself these questions when searching for answers to your research question:
      • Where would you go to find more information about your topic? 
      • Who is an expert? 
      • What types of sources would you use to answer your questions?
      • These questions will lead you toward the types of sources that you need to find. Won’t always be scholarly sources that are in databases, sometimes you’ll discover what you need is a tiktok, twitch stream, piece of misinformation, etc.