Working with ChatGPT: Research Paper Student Guide [Strategies]
10-15 minute read
Mary Landry
What You Will Learn in This Section
By the end of this tutorial, you will be able to utilize a specific formation of generative AI (GenAI)—the prominent Large Language Model (LLM) ChatGPT—as an aid within the research paper writing process to
- survey the ongoing discourse of research on a given topic
- draft with different reasoning strategies
- integrate sources and quotes
Additionally, you will critically reflect on the possible pitfalls in regards to originality and time management when using ChatGPT as an aid for composing a research paper.
What key terms should I know within artificial intelligence (AI) discourse?
Artificial Intelligence (AI)
Artificial Intelligence, often referred to as AI, encompasses computational systems created to imitate human behaviors and cognitive processes. In essence, computer scientists strive to design AI that can augment or replicate specific human cognitive functions, such as problem solving, experiential learning, pattern recognition, and decision making.
Generative AI (GenAI)
Generative AI is a branch of artificial intelligence trained on vast datasets (text, images, audio, code, etc.) in order to generate contextually relevant outputs. ChatGPT is a well-known formation of GenAI designed to generate human-like language in response to a prompt; in fact, the “G” in ChatGPT stands for “generative” to reflect this nature. However, GenAI covers a great variety of generators, such as image generators like Midjourney and DALL-E 2, or voice generators like Eleven Labs. Amid these various formations, ChatGPT belongs to a specific subcategory of GenAI called Large Language Models.
Large Language Models (LLMs)
Large Language Models are a specific subset of GenAI designed to process and generate human language. These models, trained on extensive textual data, can understand, generate, and manipulate text in ways that resemble human communication. LLMs include popular chatbots like OpenAI’s ChatGPT, Microsoft’s Bing, Google’s Bard, and more.
What should I know about composing a research paper before I begin this tutorial?
A Brief Overview
In many ways, pursuing a college degree is synonymous with doing research. Indeed, writing a research paper mirrors the practices of intellectual exploration in places like the Library of Alexandria of ancient Egypt. While the modern research paper looks much different than its ancient predecessors, its value remains the same: Research is a means for finding clarity amid a world of complexity.
With the latest GenAI, such as ChatGPT, students and researchers alike now have a powerful tool at their disposal to enrich the research process. As an illustration, this tutorial will explore three specific strategies where collaborating with ChatGPT can augment traditional research methods. For further information on research papers in general, see table 1.
Table 1
Open-Access Resources Related to Composing a Research Paper
- Writing Guide with Handbook[1]
- Chapter 11 “Reasoning Strategies”
- Chapter 12 “Argumentative Research”
- Chapter 13 “Research Process”
- Informed Arguments [2]
- Section VII “Researched Writing”
- How Arguments Work – A Guide to Writing and Analyzing Texts in College[3]
- Section 6 “The Research Process”
What are three ways ChatGPT can support research paper composition?
Surveying the Ongoing Discourse of Research on a Given Topic
No research paper is composed in a vacuum; instead, scholarly work is built upon existing knowledge, theories, and previous research findings. As such, the research process has often been likened to joining an “ongoing conversation,” as Kenneth Burke famously illustrates with his parlor metaphor.[4] This illustration can help you think critically about how your research fits into the ongoing discourse on a topic. How will you build upon what has already been said? Do you agree or disagree with others’ viewpoints? What new perspectives can you add to the dialogue? Such questions can help you lean into the scholarly exchange of ideas and position your research to be a more meaningful contribution to the collective understanding of a subject.
When researching a topic, understanding the academic conversation surrounding it can be challenging, especially for popular topics like climate change or gun control. This is where the design of tools like ChatGPT can be uniquely helpful. As an LLM, ChatGPT has been trained on vast arrays of internet text, making it well-versed in the ongoing discourse of research across various subjects. Consequently, ChatGPT can serve as a valuable entry point for surveying the most relevant language related to a research question.
For an example, see Strategy in Action #1, in which a user puts in their research question on internet privacy and ChatGPT responds with an overview of the different perspectives involved.
Strategy in Action #1
View the following example prompt and response with ChatGPT focused on surveying the ongoing discourse of research on a given topic.
If you’d like to interact directly with ChatGPT in the above conversation, access the original chat through this link[5] and click the ‘Continue the conversation’ button in the opened dialogue window.
Check out table 2 for additional tips and suggestions related to surveying the ongoing discourse of research on a given topic.
Table 2
Tips and Suggestions: Know the Limitations of ChatGPT’s Training
An essential component of familiarizing yourself with the ongoing discourse around a research topic is being up to date on the most recent developments and perspectives. For example, if someone wrote on the topic of abortion but didn’t acknowledge that Roe v. Wade had been recently overturned in June 2022, it would seem like they weren’t well-versed in their research.
In relation to this point, it’s important to acknowledge that ChatGPT-3.5 (the current “free”[6] version of the software) cannot generate output related to knowledge or events that occur after September 2021. It does not know, for instance, that Roe v. Wade has been overturned.[7] Indeed, it is important to remember that ChatGPT is not designed to be a search engine like Google; it is designed to mimic human language processing and production. In other words, it is designed to favor coherence over accuracy.
While some LLMs like Microsoft’s Bing, Google’s Bard, and ChatGPT-4 (the subscription version) can generate real-time information like the overturning of Roe vs. Wade, knowing this temporal limitation of ChatGPT-3.5 is a reminder to use any LLM as a starting point for getting a feel for the discourse around a topic. Otherwise, you may be hearing just a portion of the conversation.
Drafting with Different Reasoning Strategies
Most research paper assignments tend to feature larger page requirements to accommodate an in-depth exploration of a topic. However, the longer length can also present the challenge of sustaining a well-developed argument throughout the paper. To tackle this challenge, it is important to think critically about the most effective and persuasive way to approach a topic.
For this purpose, writers and speakers employ various reasoning strategies to build and structure an argument, such as analogy, cause-and-effect, classification and division, compare and contrast, problem-and-solution, definition, and more. Consider, for instance, how these various reasoning strategies can generate a number of ways to approach the topic of mental health, as demonstrated in Table 3:
Table 3
Various Reasoning Strategies for Approaching the Topic of Mental Health
- Analogy: You might draw parallels between self-compassion and car maintenance to highlight the importance of regular maintenance for optimal performance.
- Cause-and-Effect: You might research how social media contributes to the development of mental health issues like anxiety or low self-esteem.
- Classification and Division: You might investigate the various subtopics that contribute to mental health issues, such as biology, environment, or psychology.
- Compare and Contrast: You might analyze the similarities and differences between how mental health is handled in high schools vs. colleges.
- Problem-and-Solution: You might identify the challenges in accessing mental health care and proposing strategies to improve affordability and availability.
- Definition: You might focus on clarifying key terms related to mental health, like stigma, resilience, and mindfulness, to foster a common understanding.
If you’re feeling stuck with how any of these reasoning strategies might apply to your research, ChatGPT can be a linguistic lifeline. As an LLM, the chatbot has been extensively trained in each one of these reasoning strategies, enabling it to model how these various modes might work with your topic. See Strategy in Action #2 for an example, in which a user explores several approaches for the topic of eating disorders. Collaborating with ChatGPT in this way can help you determine what to include in your writing as well as which strategy (or mix of strategies) is the best for organizing your thinking.
Strategy in Action #2
View the following example prompt and response with ChatGPT focused on drafting with different reasoning strategies.
If you’d like to interact directly with ChatGPT in the above conversation, access the original chat through this link[8] and click the ‘Continue the conversation’ button in the opened dialogue window.
Integrating Sources and Quotes
In any research paper, the integration of sources is not just an assignment requirement to check off—it adds essential depth and credibility to your work. When you integrate sources into your paper, you are both showcasing the close reading of the research you’ve done and inviting yourself into the academic conversation. As such, the art of smoothly embedding quotes within the flow of your writing should be a key focus because it enables readers to follow your ideas seamlessly and grasp the relevance of the quotations instantly. However, this task can present a number of challenges, such as the struggle to balance your voice with the source’s voice, the fear of misinterpretation, and the need for proper contextual understanding.
To improve the integration of quotes, students can seek personalized feedback from ChatGPT, which has been extensively trained in various quoting strategies within its design as an LLM. See Strategy in Action #3 for an example, where ChatGPT generates guidance for integrating a quote from The Odyssey into the user’s analysis. ChatGPT models various quoting strategies, such as introducing the context of Odysseus’s encounter with the Cyclops, using signal phrases like “declaring,” and explaining the significance of Odysseus’s taunt. This practical example showcases how ChatGPT can help you not only create well-embedded sources, but also clearly communicate your meaning.
Strategy in Action #3
View the following example prompt and response with ChatGPT focused on integrating sources and quotes.
If you’d like to interact directly with ChatGPT in the above conversation, access the original chat through this link[9] and click the ‘Continue the conversation’ button in the opened dialogue window.
Check out table 4 for additional tips and suggestions related to integrating sources and quotes
Table 4
Tips and Suggestions: Avoid Sharing Sensitive or Confidential Information
As you work on your research paper, it’s important to carefully consider the integration of any research that involves sensitive data, such as confidential personal interviews with survivors of traumatic events. While a source like this may be valuable to include in your research paper, you should avoid sharing such confidential information with ChatGPT.
Many LLMs like ChatGPT do not necessarily protect private information and users should not assume their private information is protected; in fact, the LLM will use such information for training purposes. Because private information can become part of what is known as collective intelligence, any questions, phrases, expressions, ideas, or information you input into ChatGPT will become part of its learning source.
As such, if you need help with embedding a quote or source that features sensitive information, it is better to seek advice directly from an instructor or trusted advisor, rather than to seek personalized feedback from ChatGPT.
What’s a pitfall to avoid when collaborating with ChatGPT on a research paper?
Overreliance on ChatGPT Can Cost Originality and Time
The research paper assignment is often the longest essay you will write for a rhetoric and composition course, which makes it tempting to rely heavily on a tool like ChatGPT to slog through to the finish line. This temptation is not your fault—the brain is hardwired to prefer the path of least resistance. However, you can practice self-discipline through the knowledge that there is a price for overreliance, particularly in regard to originality and time.
Consider first the basic purpose of doing research: to make original discoveries or insights amid the complexities of the world. Overusing ChatGPT undermines this purpose. You run the risk of just echoing its output, which lacks originality since it relies on existing data patterns. It can’t make original discoveries or insights; that’s your role.
Additionally, consider how ChatGPT can affect time management. It is easy to spend a significant amount of time interacting with the chatbot without making significant progress on an assignment. Imagine, for example, a student who is researching climate change but gets caught up exploring related topics (e.g. deforestation) through their interactions with ChatGPT. While such engagement is interesting, this student could end up not having enough time to finish their work. In general, part of learning how to write a research paper is the time management the assignment requires. Research papers are too complex and too lengthy of a writing assignment to put off to the last minute. As such, relying too heavily on ChatGPT can impede finishing a research paper in a timely manner.
Learning Outcomes for ENGL 1302: How does this tutorial apply to state standards?
This tutorial is designed to support your success in ENGL 1302 (Composition II) by aligning with the student learning outcomes established by the Texas Higher Education Coordinating Board. As highlighted in Table 5 below, this tutorial directly addresses three key standards. By aligning with these specific learning outcomes, this tutorial not only provides you with skills and knowledge that transfer across diverse learning environments but also gives your education value outside your institution.
Table 5
Applicable State Learning Outcomes
- Demonstrate knowledge of individual and collaborative research processes: By utilizing ChatGPT as a tool to survey the ongoing discourse of research on a given topic, you will be engaged in collaborative research processes, working with the AI to enhance your understanding and analysis of various topics.
- Develop ideas and synthesize primary and secondary sources within focused academic arguments, including one or more research-based essays. When collaborating with ChatGPT on different reasoning strategies for drafting, you will be empowered to organize your writing and thinking into focused academic arguments.
- Write in a style that clearly communicates meaning, builds credibility, and inspires belief or action. With ChatGPT‘s assistance in modeling various quoting strategies, you will be empowered to integrate sources seamlessly into your writing, helping you to clearly communicate your meaning and further build credibility with your audience.
- Robinson, Michelle Bachelor, et al. Writing Guide with Handbook. OpenStax, 21 Dec. 2021, https://openstax.org/details/books/writing-guide Accessed 21 Sept. 2023. ↵
- Pantuso, Terri, et al. Informed Arguments: A Guide to Writing and Research. 2023. Pressbooks, https://pressbooks.library.tamu.edu/informedarguments/.Accessed 21 Sept. 2023 ↵
- Mills, Anna. How Arguments Work - A Guide to Writing and Analyzing Texts in College. Humanities LibreTexts, 6 Oct. 2019, https://human.libretexts.org/Bookshelves/Composition/Advanced_Composition/How_Arguments_Work_-_A_Guide_to_Writing_and_Analyzing_Texts_in_College_(Mills). Accessed 21 Sept. 2023. ↵
- Brent, Doug. "Burke’s 'Unending Conversation' Metaphor." Rhetorics of the Web: Implications for Teachers of Literacy, https://kairos.technorhetoric.net/2.1/features/brent/burke.htm. Accessed 21 Sept. 2023. For the original source of the metaphor, see pp. 110-111 from Kenneth Burke’s The Philosophy of Literary Form: Studies in Symbolic Action, published by Louisiana State University Press in 1941. ↵
- “Surveying the ongoing discourse of research" prompt. ChatGPT, OpenAI, 24 July 2023, https://chat.openai.com/share/db83cde1-0d70-4311-a5f4-5f62b3e4f31f. ↵
- I’ve put “free” in scare quotes to draw attention to an underlying contradiction of perceiving ChatGPT as a free service. Yes, a user doesn’t have to pay money to use ChatGPT-3.5, but they are paying OpenAI with any data they input into ChatGPT, which helps further train the model. ↵
- “Roe v. Wade status" prompt. ChatGPT, OpenAI, 1 Aug. 2023, https://chat.openai.com/share/78bec693-7da0-45ef-9399-6d37679bcdc0 ↵
- “Drafting with different reasoning strategies" prompt. ChatGPT, OpenAI, 23 July 2023, https://chat.openai.com/share/c63e8517-5fab-4862-a946-ad22321da401 ↵
- “Integrating sources and quotes" prompt. ChatGPT, OpenAI, 23 July 2023, https://chat.openai.com/share/e4356f61-b9ae-4b2f-b38b-e172d82b8f03 ↵