Harnessing OpenAI's Writing Tips: A Step-by-Step Portfolio for AI-Assisted Learning

A practical guide for educators to embed AI thoughtfully into the writing process while maintaining student learning integrity.

***This post was written with the assistance of ChatGPT***

On Wednesday November 20th, Nick Potkalitsky and I will be joined by Rob Nelson for a LinkedIn Live discussion of our recently released book; “AI in Education: A Roadmap for Teacher-Led Transformation.” You can register here. We look forward to seeing you there!

In today's rapidly evolving educational landscape, Humanities teachers are grappling with a significant shift: artificial intelligence is now capable of generating many traditional educational outputs at a level that rivals or even exceeds human performance. Faced with this new reality, educators are exploring diverse methods for integrating AI into the learning process without undermining students' critical thinking skills.

One such effort comes from OpenAI, which yesterday released "A Student's Guide to Writing with AI." This guide presents several interesting approaches to help students effectively use AI tools like ChatGPT to support their writing process without sacrificing their learning. Some might argue that these suggestions are a day late and a dollar short, but several of these ideas, in my opinion, do create space for metacognitive reflection and a deeper understanding of how to leverage AI as an educational augmentation tool rather than a replacement.

Jason Gulya, a leading advocate in this space, proposes a portfolio-based methodology that emphasizes the cognitive process over the final product by having students submit their thinking in stages. Building on this idea, I've combined Gulya's approach with OpenAI's strategies and elements of the “Grade the Chats” framework to develop a "Portfolio Approach" that integrates AI into the writing process while aiming to preserve and enrich student thinking.

As always, consider the below concepts as initial ideas. They can and should be adapted to every teacher’s needs, as well as variables of student age, content, discipline, and skill.

1. Step-by-Step Portfolio with AI-Integrated Reflection

• Idea: Use ChatGPT to support and document the essay-writing process, and have students reflect on the AI's involvement at each stage.

• Implementation:

  • Initial Research Phase: Students use ChatGPT to generate a broad overview of their topic (e.g., "Give me an overview of theories of motivation"). They include a summary of this session in their portfolio, noting key takeaways, any potential biases, and questions that arose from the user perspective.

  • Outlining and Structure Feedback: Students draft an outline and then use ChatGPT for structural feedback (e.g., "Review my outline and give me feedback on the structure"). They add the original outline, AI feedback, and their own notes on changes they accepted or rejected, explaining why.

  • Reverse Outlining for Flow: After drafting, students use reverse outlining to assess coherence (e.g., "Create a reverse outline of my essay"). This helps them evaluate if their ideas logically flow, and they reflect on how this shaped the next round of edits.

  • Submit/Share Specific Chats: At any point in this process, students may submit all or a portion of their chats for review and discussion. Chats can be used to facilitate peer feedback on both the user’s approach to AI as well as an analysis of the AI outputs. Teachers can also provide direct feedback as needed.

2. Socratic Dialogue as a Thinking Development Tool

• Idea: Encourage students to use a Socratic dialogue with ChatGPT to deepen their arguments and explore opposing viewpoints.

• Implementation:

  • Thesis Development: Students engage in a Socratic dialogue with ChatGPT to develop their thesis and identify diverse perspectives (e.g., "Interview me about my ideas on Virginia Woolf’s use of stream of consciousness"). This dialogue can be recorded in their portfolio as a transcript.

  • Reflection on AI Interaction: In their portfolio, students analyze how ChatGPT's questioning helped refine their thesis. They should note specific moments where the AI pushed their thinking, including what suggestions were adopted, why, and how their argument evolved as a result.

  • Counterargument Exploration: Students use ChatGPT to generate counterarguments to their thesis (e.g., "Challenge my thesis and provide potential counterarguments"). They document these counterarguments and their responses in the portfolio, highlighting critical revisions and reasoning.

3. AI as a Feedback Tool: Iterative Drafting and Citation Support

• Idea: Use ChatGPT to iterate on drafts and automate citation-related tasks, allowing students to focus on conceptual development and articulation.

• Implementation:

  • Draft Feedback and Refinement: Students submit early drafts to ChatGPT for iterative feedback on structure, clarity, and argument strength (e.g., "Suggest five ways I can improve my draft without rewriting it for me"). They then include both the original draft and revised version, detailing changes made and their thought process behind each in a separate reflection.

  • Citation Management: Students use ChatGPT for citation formatting and verifying inline citations (e.g., "Format my bibliography in MLA"). This stage includes a reflection detailing how the user cross-checked the AI output against citation standards and an authorial ownership statement regarding the veracity of the citations.

Outcome of the Portfolio Approach

This approach encourages students to focus not only on the final essay product but also on their learning journey. The key elements of these approaches include the submission of AI interactions for review, periodic or holistic reflections on the value and efficacy of AI use, and discussions or written explanations around how their argument or final product evolved.

By moving educators away from assessing the product of student thinking and into evaluating the process, we can maintain assessment integrity and even – potentially – deepen our student’s cognitive approaches. An increased emphasis on reflective thinking at every stage of the learning journey not only encourages metacognition but also emphasizes the need for critical review of AI outputs and interactions.

These types of learning approaches have the potential to be transformative, and they begin with the development of AI Literacy.

What are your thoughts? Share your comments and ideas below.