Breaking Down the Writing Process with AI
Breaking Down the Writing Process with AI
The instructional practices shared in this article are ideas for exploration, not requirements for any instructor. They were developed by Northwestern IT Teaching and Learning Technologies in partnership with the Provost’s Generative AI Advisory Committee. Please note:
- Accessing Copilot via your Northwestern credentials is the recommended path for accessing a generative AI tool. Questions about whether a risk assessment has been performed or an institutional contract exists for a specific AI tool can be directed to the Northwestern IT Information Security Office (security@northwestern.edu). Procurement of new AI tools should follow university processes and policies regarding licensing and third-party risk assessments.
- Output from large language models (LLMs) can include false or incorrect information. Verifying accuracy via other sources is a critical practice for instructors, students, and staff to engage in when using LLMs.
For students and instructors, the introduction of generative AI via chatbots like ChatGPT in late 2022 introduced significant and complex challenges to core aspects of teaching and learning in higher education. Currently, the Northwestern Office of the Provost supports instructors in choosing their own level in which to integrate generative AI into their courses and includes it in the University’s Principles Regarding Academic Integrity. For many, involving generative AI in skills like writing is, at best, an uncomfortable position. For some, exploring generative AI with their students or asking them to critique ChatGPT output has been a productive way to engage students in examining the impact it can have on their writing.
When we think about generative AI and writing, a great place to start is with this short video featuring Elizabeth Lenaghan, director of The Cook Family Writing Program and associate professor of instruction, where she reiterates the importance of teaching writing in the generative AI age via a process model. The Cook Family Writing Program has created specific resources on generative AI and writing, which are referenced throughout this article. This article looks at four distinct parts of the writing process and offers step-by-step instructions on how to incorporate generative AI in ways that can help grow students’ understanding of generative AI and their own writing skills.
Through the University’s Microsoft license, Northwestern students, faculty, and staff have access to Microsoft’s implementation of the GPT 4.0 large language model through Microsoft Copilot (available only through a smartphone app or internet browser). Access to Copilot is important because when you are signed in with your Northwestern account, any data you put into the chat is covered by Northwestern’s contract with Microsoft for data protection so that Microsoft does not use it for product improvement or to train their AI models. This is the closest interface to ChatGPT, but only provides data protections when signed in with a Northwestern Microsoft Account.
Growing Critical Generative AI Users
As fall quarter starts, understanding the basics of how large language models work is critical for instructors and students, regardless of the extent to which you do or do not allow students to use generative AI in your class. We recommend watching and sharing the videos created by the Center for Advancing Safety of Machine Intelligence included below to build a common language and understanding with your students about how generative AI works and its potential impact when used in their writing. To use generative AI tools well, students need to employ critical thinking, information literacy, and writing skills.
Writing Activities
Activity 1: Brainstorming with Generative AI
Description:
Many students find it challenging to select a topic to write about. AI can assist by suggesting and refining ideas, much like guidance from a friend or instructor.
- Without using generative AI, students brainstorm ideas related to a given topic, generating lists of ideas.
- Students review their lists to identify common themes and core concepts. For each core concept, they write a summary sentence explaining its significance.
- Students pick one summary to enter as a prompt in Copilot to explore new angles and extend their initial ideas. Be specific in the prompt about what kind of output you want to see. For example, "I am sharing an idea for an essay I will write. Give me a list of five points I should cover. Here's the idea: [Summary sentence]."
- Have students discuss these new perspectives in pairs, small groups, or larger group discussions to gain insights. (Source: Leon Furze)
AI Learning Objectives:
Students will:
- Generate and refine ideas and articulate core concepts.
- Use generative AI tools to enhance their understanding.
- Engage in discussion to develop critical thinking and communication skills.
- Begin to understand limits of generative AI LLM tools.
Additional Resource: Brainstorming with (and without AI)
Activity 2: Crafting Thesis Statements with AI
Description:
A clear thesis statement is crucial in writing as it provides direction and focus, guiding the structure and content of the entire piece. While creating a thesis can be challenging, generative AI can assist by offering suggestions and alternatives to help refine and articulate a compelling argument.
- Students start by selecting a topic related to the course content that interests them.
- Students draft a thesis statement focusing on their main argument and its significance.
- Pair students to share their statements, providing feedback to each other on clarity, argument strength, and potential improvements, with constructive criticism and specific suggestions.
- Have students input their thesis statements into a generative AI tool using the prompt, "I wrote a thesis statement, and I want you to provide me with feedback on clarity, argument strength, and potential improvements to the statement. Here is my thesis statement: [THESIS STATEMENT]"
- Have students compare the AI's suggestions with their partner's feedback, noting any unique differences.
- Conclude with a class discussion on their experiences, exploring the value of human versus AI feedback in enhancing writing. (Source: Crystal Camargo)
AI Learning Objectives:
Students will:
- Refine their thesis statements by integrating feedback from peers and generative AI.
- Improve their ability to critically evaluate and enhance their arguments.
Additional Resource: Creating a Thesis statement with (and without AI)
Activity 3: Reverse Outlining with AI
Description:
Outlining is a critical step in the writing process that helps students understand the structure and flow of their work. By using a reverse outline with traditional outlining methods and generative AI tools, students can ensure that each paragraph contributes effectively to the overall argument.
- Students select one of their own previously written papers or essays and review it to understand its main points and overall structure.
- Students create a reverse outline by condensing each paragraph into a brief statement capturing its main idea, considering how it contributes to the paper's overall argument or narrative.
- Next, students use a generative AI tool to produce a reverse outline of the same paper and compare it with their own to discover any differences or similarities. Use the prompt, “Create a reverse outline of this document that includes a one- to two- sentence summary of each paragraph. Each summary should include the main idea of the paragraph and how it contributes to the paper's overall argument or narrative. Here is the document: [DOCUMENT TEXT]"
- Encourage reflection on whether the AI highlights points they missed or suggests a different structure and how both outlines align or not with the original intentions for their paper.
- Facilitate a class discussion where students share insights gained. (Source: Ohio College Teaching Consortium)
AI Learning Objectives:
Students will:
- Use AI tools to create reverse outlines of their work, comparing AI interpretations with their own.
- Improve their understanding of structure and clarity in writing.
Additional Resource: Outlining with (and without AI)
Activity 4: Draft Feedback with AI Integration
Description:
Students can utilize AI to "read" their draft and receive feedback on missing components, potential counterarguments, and structural improvements.
- Ask students to bring a draft of a paragraph or two to class for feedback.
- Students exchange drafts in small groups, providing and receiving peer feedback focused on structure, clarity, and argument strength.
- Students input their draft into Copilot to receive additional feedback.
- Students compare Copilot’s suggestions with the feedback they received from peers and the instructor.
- After reviewing insights from all sources, students can develop a revision plan.
- Conclude with a class discussion on how AI can complement traditional feedback methods, enhancing students' understanding of their strengths and areas for improvements.
(Source: Ethan Mollick and Lilach Mollick)
AI Learning Objectives:
Students will:
- Integrate feedback from peers, instructors, and generative AI to improve their writing.
- Enhance clarity, coherence, and argument strength in their drafts.
- Understand the role of AI in the feedback process.
Final Thoughts
When exploring generative AI in your courses, keep the generative AI portion opt-in and share with students how they complete the activity without using generative AI. No matter if or how you utilize generative AI in your course, be explicit with your students about your policies and expectations.
Connect with your community through colleagues, Northwestern writing experts, or request a consultation to talk through any questions you have about using generative AI in the writing process.
Videos to Share with Students
This video describes how LLMs use predictions to create output and how hallucinations can occur.
This video outlines what’s happening when it looks like LLMs are “thinking.”
From Northwestern Center for Advancing Safety of Machine Intelligence.