How learners should engage Large Language Models framework
Framework Justification: Mollick & Mollick (2023)
Core Thesis
This analytic framework builds directly on Mollick & Mollick’s (2023) “Assigning AI: Seven Approaches for Students, with Prompts.” Their work outlines how students should engage Large Language Models (LLMs) through intentional prompts, explicit oversight, structured reflection, and transparent integration.
I adapted their pedagogical roles (Tutor, Coach, Mentor, etc.) into a research lens—shifting the focus from “how students should use AI” to “how students actually used AI in authentic literacy inquiry.”
Theoretical Contribution (Adaptation)
Where Mollick & Mollick provide a pedagogical scaffold for classroom practice, this project turns those scaffolds into a qualitative analytic lens, allowing us to examine how preservice literacy educators actually enacted these roles and responsibilities and negotiated Boundary-work and Epistemic Stance in authentic inquiry.
The adaptation transforms their guidance into:
- An analytic tool for qualitative coding.
- A lens on real student interactions and artifacts.
- A model for studying AI-mediated literacy practices.
Direct Connections to Source Text
The framework's structure is a direct methodological response to key Mollick & Mollick guidelines:
| Framework Component | Direct M&M Connection | Key Quote / Concept |
|---|---|---|
| Inputs | Setting the stage for the AI’s function. | “Provide context and goals to the AI.” (M&M, various prompts) |
| Prompts | The core mechanism for defining the interaction. | “Assign roles” & shape the AI’s function (e.g., AI-Tutor, AI-Coach). |
| Outputs | The need for vigilance and assessment. | “Avoid complacency about AI responses.” Emphasis on hallucination and bias risks. |
| Integration | The requirement for synthesis and judgment. | “Students must remain the human in the loop.” Avoid outsourcing analysis. |
| Reflection | The crucial step for learning and ethical development. | Explicit requirement to “Prompt metacognition” (AI as Coach) and articulate takeaways. |
. Inputs – What they upload
This aligns with the essential action of Prompting discussed throughout the paper. The quality and nature of the input heavily influence the AI's utility. The paper highlights the importance of providing context and clear goals when students are engaging the AI in roles like AI Mentor or AI Tutor.
- Source Connection: The entire paper revolves around prompts—the inputs students use to assign roles and tasks to the AI. For example, the detailed prompts for AI as Mentor instruct students to clearly articulate their goals and learning level as part of the initial input.
2. Prompts – How they engage the AI
This is a central concept in the source material, as the paper is subtitled "with Prompts." This part of your framework captures the student's agency and strategy in directing the AI.
- Source Connection: The paper dedicates sections to "Example Prompts" for all seven approaches (Tutor, Coach, Mentor, etc.), underscoring the critical role of the prompt in shaping the learning experience and output.
3. Outputs – What the model produces
This directly corresponds to the AI's response and is a core component that the students must critically assess according to the authors.
- Source Connection: The document repeatedly discusses the risks associated with the AI's output, such as Confabulation (hallucination) and Bias Risks. Students are challenged to remain the "human in the loop" and not develop "complacency about the AI’s output," making the nature of the Output a crucial point of analysis.
4. Integration – How outputs shape the final project
This step analyzes the Boundary-work between the student's original effort and the AI's contribution. It moves beyond simple output analysis to examine the ultimate application and synthesis of the AI-generated material.
- Source Connection: The paper's guidelines consistently push students to "actively oversee the AIs output, check with reliable sources, and complement any AI output with their unique perspectives and insights." This process of complementing and integrating the AI's work into the final student product is exactly what this step captures.
5. Reflections – Metacognitive and ethical insights
This is strongly supported by the paper's focus on metacognition and responsibility. It directly addresses the epistemic stance (how students view knowledge and their role in creating it) and ethical agency.
-
Source Connection:
-
Metacognition: The AI as Coach approach is explicitly defined as "Prompt metacognition" with the pedagogical benefit of providing "Opportunities for reflection and regulation, which improve learning outcomes." The reflection prompts provided in the paper are designed to make students articulate their thoughts about the experience.
-
Ethical/Accountability: The student guidelines for the AI as Mentor state: "Share your complete interactions with the AI... What are some of your takeaways in working with the AI?" and explicitly warn about privacy risks, emphasizing accountability and ethical use.
-
💡 Highlighting Core Concepts
Your framework effectively captures the three core concepts you mentioned by breaking down the student-AI interaction:
-
Agency: Highlighted in Inputs and Prompts, where the student actively makes choices to direct the AI.
-
Boundary-work: Captured in Integration, where the student separates and combines their own work with the AI's output.
-
Epistemic Stance: Explored in Reflections and the critical evaluation of Outputs, where the student determines the trustworthiness and value of the AI's "knowledge."