Clarifying AI Use in Academia: How to Create Use Cases that Call upon the Strengths of Multiple AI Resources
Using generative artificial intelligence (AI) tools for academic purposes is not a futuristic notion; it's a present-day reality.
Much of our work as early adopters of AI in academia relates to helping clarify how AI can be used ethically and responsibly – we repeatedly reiterate the importance of functional (basic use), critical, and rhetorical AI literacy.
Moxie’s AI Literacies Framework, based on the work of Selber (2004).
This blog post falls squarely under functional AI literacy, which we define as understanding the basics of how AI systems and their underlying technologies work. We hope to help you better navigate AI interfaces and platforms and identify (and clarify) limitations and potential in various models. We hope that you will be able to better understand the various academic use cases for premium AI models like Gemini Pro, GPT-4, and Claude-3.
Choosing Appropriate Models: The Literature Review Use Case
Selecting the appropriate AI tool depends on what you want to use it for - the specific use case. While many of the top AI models available (e.g., GPT-4, Gemini Pro, Claude 3) can produce natural-sounding text and maintain context through dialogue, there are scenarios for when you might choose one AI model over another.
To give you a sense of how these models might be used for various academic use cases, review the chart below, which compares top AI models against academic use cases. As you do, consider which use cases apply to you in your work.
Table of Use Cases Matched to Generative AI Models.
Based on the above chart and your needs, who is your winner? Claude3? GPT4? Gemini or Perplexity? You likely discovered that no one model can meet your needs, yet you might not know when and how to bring a different model to bear.
That’s why we created a scenario around one of the more complex projects an academic writer often needs to complete:
The literature review.
Sure, you can just dive in and ask a chatbot for help, but you can also plan for a literature review workflow that incorporates the strengths of different models and tools. Below, we outline one approach for using a combination of models to help literature review writers. We map tools/models to common literature review tasks so that you can see how this use case work powered by multiple AI models might look.
Search for relevant literature with our literature review tool.
To search the literature, literature review readers need to use relevant search terms that are appropriate to their research topics. Our Literature Review AI Tool will help generate suggestions for keywords, synonyms, search strings, and Boolean operators to help you identify relevant literature. You can also use this tool to analyze a recently published article to help you solidify a search strategy or identify additional concepts and keywords to broaden your search.
Fact check and search the Internet with Gemini or Perplexity.
Once you have your search terms, you might transition to Perplexity or Gemini Pro, both connected to the internet, to locate open-access sources or up-to-date information on your research topic. Gemini scans databases for scholarly articles, applying advanced algorithms to sift through data. Perplexity is an alternative to a traditional search engine but uses Open AI models to provide answers linked to sources on the internet. Note: Utilizing your critical AI literacy as you vet sources is essential.
Summarize research articles and consider theoretical perspectives with GPT-4 Turbo or Claude-3 Opus.
You can then transition to GPT-4 to brainstorm and ideate complex concepts that require more advanced reasoning skills or take advantage of its large context window to summarize research articles. For example, you might use GPT-4 to explore various theoretical perspectives and reason through theoretical implications to solidify your research framework.
Organize and synthesize literature findings with our literature review tool.
AI can categorize articles based on their methodology, findings, or topic. This can help you to identify significant trends and themes within the research literature. Once this categorization is complete, you can utilize our Literature Review AI tool to analyze a literature matrix that includes information about relevant articles on your topic. This does not replace the work you do looking for themes in your literature. Still, the tools can help you see trends and patterns in your literature review methodology, setting, and finding — a critical synthesis step. Since a vital marker of a well-written literature review is identifying gaps, you have to understand the trends in the literature, and our tool can provide you with top-notch assistance.
Get feedback on your literature review with our literature review tool.
Once you have written a draft, use our Academic Writing Tools to get instant, high-quality feedback on your literature review draft. This AI tool can suggest structure improvements, identify content gaps, and ensure that your argumentation is coherent and aligned with your research objectives. The power of this tool
Disseminate your research and establish thought leadership with Claude.
When preparing to publish, you might use Claude’s friendly and conversational tone to craft a draft of a personal cover letter for your manuscript or a LinkedIn post to share your research after publishing.
The above use case represents just one way that we can bring multiple models and tools to bear on a project without sacrificing the academic integrity of our work.
Planning your Use Case
We created the above academic workflow with several AI models because we understand two critical areas of this project: (1) knowledge of how to write a literature review and (2) the capabilities of the models and tools available to us.
Try the following workflow to determine which models and tools you can use to create a workflow for an academic research project.
Give your project a name (e.g., “Methodology Chapter”).
Research the types of research/writing tasks you need to be able to perform. If you need assistance, our new all-in-one research and writing tool can help!
Arrange your tasks into a workflow. Think in phases of work.
Pair your tasks with an appropriate AI model or tool. Use the table above or our tools to assist you with the key phases of your use case.
The more clarity we gain around what we are trying to accomplish as research writers, the more we can make informed decisions about how to bundle AI usage, calling upon the strengths of the various platforms available. For this reason, we have created a membership that gives you access to several platforms and our suite of prompt-engineered AI research and writing tools. We hope you can design and execute use cases that serve your work well.
