Research is a process of discovering information, evaluating that information, and reshaping our understanding of a topic or issue as we search. It requires basic search mechanics as well as higher order thinking such as evaluating and creating.
AI offers us the opportunity to consider the research skills and practices that we want students to acquire in our classrooms and whether AI will be a part of that research process.
AI and other research tools will constantly be evolving, but we can provide students with the critical thinking skills needed to evaluate those tools and their content.
The threshold concept, Research as Inquiry, stresses that research is iterative and involves asking new questions and breaking down complex questions into simpler ones. As students become more information literate, they develop the ability to:
Students can achieve these dispositions through certain knowledge practices like: formulating research questions, simplifying topics and concepts, generating keywords and outlines for projects, organizing information, finding a variety of sources, and critiquing or analyzing content.
People are now using generative AI to complete some of these tasks. AI may shift how we help students reach these critical information mindsets by practicing new research processes and skills.
For example: we value that the research process is iterative and non-linear because we want students to develop mental flexibility. When students learn to search library databases, they practice skills like refining research questions, brainstorming search terms, assessing the quality of search results and refining search queries. As an instructor, you might ask how prompt engineering and assessing Generative AI output could reinforce mental flexibility in similar ways. In both cases students have to define their information need and engage in metacognition about their search process.
Using AI, Spotify generates personalized playlists ("daylists") with word-salad titles based on past listening sessions. The playlists are a great way to "discover" new music that is similar to past music that you've listened to. But using AI for discovery tools raises important questions about relevance, ranking, and selection.
Personalization of music playlists is the tip of the AI iceberg - we know major library tools provided by vendors like ESBCO and JSTOR are beginning to incorporate AI into the discovery and search process. Library tools are shifting how they present and rank sources based on additional algorithms running behind the scenes. AI tools like Elicit and Reseach Rabbit will create a web of related research for scholars.
These new tools and features will change how we interact with database search result and influence how and what we find in traditional library spaces.
Research Process
Because technology is constantly changing, we can focus on the process and outcomes more than the tools themselves.
Research AI Tools
In relation to the questions above -
Curiosity and Non-Linear Research
Instructors have concerns that AI will flatten the information landscape and students will just be "fact-finders" instead of active thinkers in the research process.
1. What AI tools have you used in the past for classwork or paper assignments? Was there a specific process that you used it for? Why? Did you find it valuable and useful?
2. What parts of the research process do you struggle with? Why? What parts are easier for you? Why? What strategies do you use to be successful? What does your research process look like normally?
3. How do you evaluate information? How do you "fact-check," or determine the credibility of a source? What are important aspects to consider? [This can change by discipline!]
4. Library databases like JSTOR and EBSCO are using AI to generate summaries or rank and promote "relevant" sources. How could this selection potentially be a discovery issue if we only see results from heavily cited scholars and only a few sources from new scholars? What perspectives are lost when only the dominant literature is retrieved?
5. Research tools like Elicit and Research Rabbit, built on large language models, are introducing new ways of finding research articles and visualizing connections between articles. The creators at Elicit even note that there are limitations to their tools and outline how to use it responsibly. How do we balance using these new tools with known limitations with the exciting advantages that they have? What critical mindsets do we need to have when using new tools built on rapidly changing technology? Is searching a skill that should be automated by a tool? When might human interaction be necessary for a search?