Skip to Main Content
Library.MacEwan.ca

Writing Guides

MacEwan University Writing Centre

The GenAI Writing Process

Pigg has found that "requesting, evaluating, and refining" are "three categories of practice that describe research writers' interactions with generative AI across research design, writing research genres, and proofreading and editing research communication" (2024, p. 2). The following guide adapts Pigg's findings to suggest a three-part process for academic writing with GenAI. 

Requesting Text

Prompting and Questioning 🤔

When using GenAI to generate text, prompt engineering--how you frame your requests for information--directly affects the quality of the output. 

📝 For pre-writing

In the pre-writing stage, GenAI can help you to: 

🔍 For secondary research

GenAI has real limitations as a search engine for scholarly research. It cannot conduct sophisticated searches or access paywalled content or non-digitized sources. Rather than admitting to its limitations, moreover, it often resorts to making things up (i.e., hallucinating). Independent research through MacEwan Library or platforms like Google Scholar remains the best way to collect secondary research on your topic.

However, Gen AI can help you to: 

  • Summarize and/or organize complex ideas or articles
  • Explain difficult concepts in simpler terms
  • Brainstorm potential research starting points on a specific topic
  • Request synthesis of multiple sources

✍️ For drafting help

In the drafting and revising stage, you can request assistance with: 

  • Organizing paragraphs or sections (introductions, conclusions, etc.) and transitions 
  • Meeting all assignment criteria 
  • Adhering to genre conventions or typical structures  
  • Exploring possible titles 
  • Formatting in-text citations and reference lists 

✂️ For editing assistance

Try asking GenAI for help with: 

  • Rephrasing text for clarity, perhaps focusing on phrasing or tone 
  • Shortening text to meet word counts
  • Proofreading the text (e.g. checking grammar or spelling) 

💡 Tip: Be specific in your requests. Instead of "Help me plan my research paper," try "Create an outline for a 5-page research paper on the impact of social media on teenage mental health."

Evaluating Text

Critical Assessment 🧐

This critical step is where your own expertise and judgment come into play. When evaluating the outputs of GenAI: 

  • Check for accuracy: Does the generated content align with your course materials and what you know about the topic? How do you know its claims are true? 
  • Assess relevance: Does the content address your specific prompt and/or research question? 
  • Verify citations: If GenAI provides references, check if they exist and say what it claims they do 
  • Consider coherence: Is there a coherent argument or point of view being presented? If so, how does that affect the presentation of the information? 
  • Evaluate depth: Is the content sufficiently nuanced and detailed?

💡 Tip: Remember, as a researcher, you should always maintain a critical stance. Don't assume GenAI is providing accurate information--it can "hallucinate" or generate false content confidently.

Refining Text

Improving the Output 🤩

The refining stage is interactive and crucial for high-quality results.

  • Follow-up questions: Ask the GenAI to elaborate on specific points that need development
  • Provide examples: Share examples of the writing style or genre you're aiming for
  • Context enrichment: Providing more information about your specific assignment will help ChatGPT tailor its responses
  • Iterative prompting: Use the AI's initial output as the basis for new, more specific prompts
  • Request alternatives: Ask for different approaches or perspectives on the same question

💡 Tip: Try targeted prompts that are based in your own assessment of the output, such as "Can you make the tone of this piece more consistent?" or "Please revise this to address inconsistent APA formatting" to refine the content. 

Examples of the AI Writing Process

Click through the tabs for examples of how GenAI can support the writing process.

🤔 Initial Request

"Help me come up with a research topic about climate change."

🤖 AI Response

Climate change is a broad topic. Here are some potential research questions:

  • How does climate change affect migration patterns?
  • What are the economic impacts on coastal communities?
  • How effective are carbon offset programs?

🧐 Student Evaluation

These questions are too broad and generic. I need something more specific that connects to my environmental science course and my interest in local solutions.

✍️ Refined Request

"I'm taking an environmental science course and I'm interested in local climate solutions. I live in Vancouver and I'm particularly interested in how urban planning might address climate change impacts. Can you help me narrow down a research question?"

💡 Improved AI Response

Based on your interest in Vancouver urban planning and climate solutions:

  • How can Vancouver's green roof initiatives reduce urban heat island effects?
  • What impact have Vancouver's bike lane expansions had on reducing transportation emissions?
  • How do Vancouver's shoreline protection strategies compare to other coastal cities facing sea level rise?
Resource: For Vancouver-specific climate adaptation strategies, see: City of Vancouver. (2022). Climate Change Adaptation Strategy. Retrieved from https://vancouver.ca/green-vancouver/climate-change-adaptation-strategy.aspx

🤔 Initial Request

"Summarize this article about quantum computing."

🤖 AI Response

Quantum computing uses quantum mechanics principles like superposition and entanglement. Unlike classical computers using bits (0 or 1), quantum computers use qubits that can exist in multiple states simultaneously, allowing faster calculations for certain problems. Current challenges include error rates and maintaining quantum coherence.

🧐 Student Evaluation

This summary is too generic and doesn't tell me anything about the specific article I'm asking about. I need to be more specific about what article and what aspects I want to focus on.

✍️ Refined Request

"I'm reading 'Quantum Supremacy Using a Programmable Superconducting Processor' by Arute et al. (2019). Can you explain their methodology for demonstrating quantum supremacy and the significance of their 53-qubit processor in terms an undergraduate physics student would understand?"

💡 Improved AI Response

Google's 2019 paper demonstrated that their 53-qubit Sycamore processor performed a specific calculation in 200 seconds that would take the world's most powerful supercomputer approximately 10,000 years.

Their methodology involved:

  1. Creating a random quantum circuit sampling problem
  2. Running this on their Sycamore processor and collecting outputs
  3. Verifying against classical simulations for smaller versions

The significance: This was the first clear demonstration that quantum computers can solve certain problems exponentially faster than classical computers—similar to the Wright brothers' first flight—brief and limited, but proof the concept works.

Citation: Arute, F., Arya, K., Babbush, R., et al. (2019). Quantum supremacy using a programmable superconducting processor. Nature, 574(7779), 505-510.

🤔 Initial Request

"Write an introduction for my essay on social media addiction."

🤖 AI Response

Social media addiction has become a prevalent issue in modern society. Platforms like Facebook, Instagram, and TikTok have captured the attention of billions of users worldwide, leading to concerns about excessive usage and dependency. This essay will explore the causes, effects, and potential solutions to social media addiction.

🧐 Student Evaluation

This introduction is too generic and doesn't reflect my specific viewpoint or research. It doesn't mention my focus on teenagers or incorporate any of the sources I've found.

✍️ Refined Request

"I need help structuring an introduction for my psychology paper on social media addiction among teenagers. I want to focus on how algorithmic design exploits adolescent brain development. I've found research by Dr. Jonathan Haidt showing correlations between social media use and depression rates, and studies by the Stanford Social Media Lab on dopamine response. My thesis argues that platforms should be regulated similar to age restrictions on other addictive products."

💡 Improved AI Response

In an era where the average teenager spends over six hours daily on social media platforms, concerns about digital addiction have moved from parental grievances to serious psychological inquiry. Recent research suggests that this addiction is not merely a matter of self-control, but a product of deliberate algorithmic design intersecting with vulnerable adolescent neurodevelopment.

The teenage brain—with its developing prefrontal cortex and heightened sensitivity to social rewards—creates a perfect neurological target for platforms engineered to maximize engagement. As Dr. Jonathan Haidt's research demonstrates, there is a concerning relationship between heavy social media use and depression, particularly among teenage girls.

This paper argues that social media platforms employ sophisticated psychological targeting that exploits adolescent neurological vulnerabilities, and therefore should be subject to age-restriction regulations comparable to those governing other addictive products.

Citation: Haidt, J., Twenge, J. M., Lozano, J., & Cummins, K. M. (2022). Specification curve analysis shows that social media use is linked to poor mental health, especially among girls. Acta Psychologica, 224, 103562.

🤔 Initial Request

"I'm writing about passive solar design in architecture. Can you edit this paragraph?"

Passive solar design uses sun's energy for heating and cooling by exposure. Sunlight makes building materials reflect, transmit, or absorb radiation. Heat causes predictable air movement. Passive systems are simple with few moving parts, require minimal maintenance and require no mechanical systems. This uses thermal mass materials, natural air circulation spaces, and building orientation.

🤖 AI Response

Passive solar design harnesses sun's energy for heating and cooling through direct exposure. When sunlight strikes a building, materials can reflect, transmit, or absorb radiation. Heat generates predictable air movement. These systems feature simplicity, minimal moving parts, and low maintenance, eliminating mechanical systems. The approach incorporates thermal mass, natural air circulation, and strategic orientation.

🧐 Student Evaluation

This edit is cleaner but still has issues: redundancy between "minimal maintenance" and "no mechanical systems," no clear connections between principles and effects, reads like a list not a system, and lacks necessary citations.

✍️ Refined Request

"I need a substantial revision that: eliminates redundancies, clarifies cause-effect relationships, creates a cohesive explanation of the integrated system, and incorporates citations from Balcomb on thermal mass and recent research on building orientation. I need an analytical tone."

💡 Improved AI Response

Passive solar design strategically harnesses solar radiation to regulate interior thermal conditions without mechanical intervention. The effectiveness depends on three integrated mechanisms: absorption of radiation by high thermal mass materials, which, as Balcomb demonstrated, can reduce temperature fluctuations significantly through heat storage and delayed release; controlled transmission through strategic fenestration; and convective air movement from temperature differentials. Liu et al. established that optimizing building envelope dimensions, particularly orientation and window-to-wall ratio, substantially enhances energy efficiency, especially when aligned within 15° of true south in northern hemisphere applications. This system leverages thermodynamic principles rather than mechanical complexity, ensuring sustainable performance with minimal maintenance.
Citations: Balcomb, J.D. (1983). "Heat storage and distribution inside passive-solar buildings." Los Alamos National Laboratory Technical Report LA-9694-MS; Liu, Z., Wu, D., Li, J., Yu, H., & He, B. (2019). "Optimizing building envelope dimensions for passive solar houses in the Qinghai-Tibetan region: window to wall ratio and depth of sunspace." Energy and Buildings, 193, 1-14.

🤔 Initial Request

"Explain intersectionality."

🤖 AI Response

Intersectionality is a theoretical framework describing how aspects of a person's social and political identities combine to create unique modes of discrimination or privilege. Developed by Kimberlé Crenshaw in 1989, it highlights how people in multiple marginalized groups experience overlapping systems of oppression.

Citation: Kimberlé Crenshaw coined the term "intersectionality" in her 1989 paper "Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics" (Crenshaw, 1989).

🧐 Student Evaluation

This is technically correct but abstract. I need concrete examples relevant to my sociology class and how researchers apply this concept.

✍️ Refined Request

"I'm writing a sociology paper on how intersectionality is applied in research methodologies. Explain with specific examples how researchers operationalize intersectionality when studying health disparities. Include scholars beyond Crenshaw."

💡 Improved AI Response

In sociology research, intersectionality becomes a methodological tool operationalized in several ways when studying health disparities:

  1. Multilevel Analysis: Dr. Lisa Bowleg advocates for examining how individual experiences connect to structural inequalities. Example: A study on diabetes management might analyze how a low-income Latina woman's health outcomes relate to neighborhood resources, healthcare access, and workplace policies.
  2. Quantitative Methods: Researchers use statistical interaction terms to capture how race, gender, and class variables interact rather than controlling for them independently.
  3. Qualitative Approaches: Patricia Hill Collins encourages phenomenological methods centering lived experiences of multiply-marginalized individuals, such as narrative interviews with LGBTQ+ veterans to understand how intersecting identities shape healthcare experiences.
Citations: Bowleg, L. (2012). The problem with the phrase "women and minorities": Intersectionality, an important theoretical framework for public health. American Journal of Public Health, 102(7), 1267-1273; Collins, P. H. (2019). Intersectionality as Critical Social Theory. Durham, NC: Duke University Press.

One final thought...

🎓 Learning to write and think for yourself 

GenAI is effective as a thought partner and drafting tool, not as a substitute researcher or writer. It is not capable of producing complex, research-based documents in your authentic voice. It cannot replace the writing process entirely, which is a good thing, since a key goal of any university education is to develop your own critical thinking, research, and writing skills.

AI tools are a powerful new support--but not a replacement--for your own intellectual development. Use them accordingly and always prioritize your own learning.

References

Pigg, S. (2024). Research writing with ChaptGPT: A descriptive embodied practice framework. Computers and Composition, 71, 102830. https://doi.org/10.1016/j.compcom.2024.102830 

Licensed under CC BY-NC | Details and Exceptions