Academic Collaboration and AI Use Framework
The Academic Collaboration and AI Use (ACAU) framework supports assignment-level decisions in courses of any modality. It does not prescribe a single approach to collaboration or generative AI. Instead, it is designed to provide a consistent vocabulary that instructors can adapt to their learning outcomes, discipline, students, and assessment context.
Why use the framework?
Students increasingly complete coursework with assistance from classmates, tutors, writing tools, generative AI, and other resources. Course-wide syllabus statements provide an important baseline, but a single statement may not offer enough specificity for every assignment, and it cannot help much if it remains buried in your syllabus. Broad directions such as “work as a group” or “AI is prohibited” often leave important questions unanswered: What assistance is permitted? What contributions are prohibited? What must students complete independently?
The framework supplements syllabus guidance with concise, assignment-specific labels. Each label helps students understand:
- how other people may contribute;
- how generative AI may contribute; and
- what evidence of their process, judgment, or understanding they must provide.
The labels do not replace complete assignment directions. Instructors should pair each label with a brief explanation of the particular permissions, limitations, tools, and requirements that apply.
What does a label communicate?
Each label combines three components.
Human Collaboration: H0–H3
How may other people contribute?
AI Participation: A0–A2
How may generative AI contribute?
Accountability Evidence: D, P, V, and E
What must the student submit or demonstrate?
A completed sequence might look like this: H2 · A1 · D · V
This label indicates that students may exchange feedback with other people (H2), use AI as an advisor (A1), disclose the assistance they received (D), and verify consequential information in their work (V).
The assignment directions should then identify the specific forms of feedback and AI support that are permitted, what students must disclose, and what content they are expected to verify.
Human Collaboration
The H code identifies the maximum role other people may play.
- H0 — Work Independently: Students complete the assignment without assistance from another person, except for George Mason’s available academic support services.
- H1 — Discuss Ideas: Students may discuss concepts, questions, and possible approaches with others. They may not share work in progress or completed answers and must complete their own work.
- H2 — Exchange Feedback: Students may share work in progress and receive comments or suggestions. Other people may not create, rewrite, or directly revise the work.
- H3 — Create Together: Students may jointly develop the work with partners or a group authorized by the instructor. This does not automatically permit collaboration with people outside that group.
AI Participation
The A code identifies the maximum role generative AI may play.
Generative AI may include standalone or embedded tools that generate or substantially revise text, images, code, analysis, or other content. Examples include ChatGPT and GitHub Copilot, as well as generative features within applications such as Grammarly, Microsoft Editor, Google Docs, and Notion.
- A0 — No Generative AI Use: Students may not use generative AI to complete any part of the assignment unless the instructor identifies a specific exception. Basic, non-generative functions, such as spelling or grammar checks, remain permitted unless the assignment states otherwise.
- A1 — AI as an Advisor: Students may use AI for specified support activities, such as brainstorming, explanation, practice, or feedback. AI-generated or substantially AI-revised material may not appear; students may not copy, closely paraphrase, or adapt AI-generated material for the submission.
- A2 — AI as a Contributor: Students may include or adapt AI-generated text, images, code, analysis, or other material, subject to the assignment’s conditions for tools, uses, disclosure, and other accountability evidence.
A2 permits AI contribution; it does not necessarily require students to use AI. If AI use is required, the assignment should state that requirement and provide an appropriate alternative when necessary for accessibility, privacy, or equitable access.
Accountability Evidence
The final letters identify evidence students must provide to make their process, judgment, or understanding visible. An assignment may include one or more evidence requirements, or none.
D — Disclose Use: Students identify how other people or AI tools contributed to their work and how that assistance influenced the submission. Disclosure creates transparency but is not, by itself, proof that all requirements were followed.
In practice, this could involve a brief statement naming a peer, tutor, or AI tool; describing the assistance received; and explaining what the student changed or chose not to change, as a result.
P — Provide Process Evidence: Students submit selected evidence showing how the work developed. This might include drafts, feedback, revision history, or influential AI prompts and outputs. Instructors should identify which artifacts are meaningful rather than requiring exhaustive documentation.
Possible evidence includes an early and final draft, selected AI prompts and outputs that materially influenced the work, or feedback and revision history accompanied by a brief explanation of how they were used.
V — Verify Content: Students independently check consequential claims, sources, calculations, code, or other outputs using an appropriate disciplinary source or method. Citation alone is not sufficient; students should identify what they checked and any corrections they made.
Depending on the discipline, students might compare a summary with the original source, recalculate a result using an accepted method, test code against known cases, or check data against an authoritative source, and identify any errors they corrected.
E — Explain Your Work: Students demonstrate their individual understanding of the work and the decisions behind it. The explanation may be written, recorded, presented orally, or demonstrated in another format.
Individual understanding could be demonstrated through a written rationale, annotated explanation, response to targeted questions, brief recording, oral explanation, or demonstration of the method or skill used.
How to apply the framework
1. Begin with the learning: Identify the knowledge, skill, judgment, or performance the assignment is intended to assess. Consider what students must be able to demonstrate themselves and which forms of assistance would support, or interfere with, that learning.
2. Select the H and A levels: Choose the maximum role that other people and generative AI may play. Select the level based on what students may actually do, not simply the names of the tools they may use. A single application may support non-generative editing, AI-generated feedback, and direct content generation.
3. Choose meaningful accountability evidence: Add D, P, V, or E when making the student’s process, verification, or understanding visible would support the learning outcomes or strengthen the credibility of the assessment. Not every assignment needs every form of evidence.
4. Add assignment-specific directions: Explain what the label means for this assignment. Identify permitted and prohibited activities, authorized collaborators or tools, required evidence, and any components students must complete independently. If accountability evidence will be evaluated, include the relevant expectations in the assignment directions and grading criteria.
The H and A codes communicate permission boundaries; they are not grading criteria. Do not grade students for the level of assistance that was permitted.
Grade the assignment’s learning outcomes and, when appropriate, the quality of evidence required through D, P, V, or E. For example, an instructor might evaluate the completeness of a disclosure, the usefulness of selected process evidence, the quality of verification, or the student’s explanation of key decisions.
If accountability evidence affects the grade, state the criteria in the assignment directions or rubric. Failure to follow clearly communicated permission boundaries may be addressed through the assignment requirements and Mason’s Academic Standards processes.
Research Brief — H2 · A1 · D · V
You may exchange feedback on drafts with peers and use generative AI to evaluate the clarity, organization, and research that supports your argument (e.g., asking ChatGPT to identify unclear sections or using Grammarly’s grammar feedback). Other people and AI tools may not write, rewrite, or directly revise your brief (e.g., copy-pasting content directly generated by ChatGPT or accepting a generative rewrite from Grammarly).
Include a 50-100-word disclosure identifying who or what assisted you, the assistance provided, and how it influenced your work; if none, state that explicitly. Also submit a table identifying three consequential claims or sources, how you verified each one, and any corrections you made.
Marketing Campaign — H3 · A2 · D · P
Your assigned team may jointly develop the campaign and incorporate AI-generated or AI-revised written and visual content (e.g., campaign copy drafted with ChatGPT or visuals created with Canva MagicDesign). Peers outside the assigned team may not contribute to the work.
Submit a contribution table identifying each team member’s contributions and each AI tool’s specific role. Include two or three outputs that most influenced the campaign, annotating what the team retained, revised, or rejected and why.
Final Exam — H0 · A0
Complete the exam independently without assistance from another person or a generative AI tool or feature. You may use only the calculator available in Honorlock and scratch paper. The exam will be proctored using Honorlock, which will monitor the session via webcam, screen recording, and automated integrity checks.
Flagged activity will be reviewed by your instructor and, if needed, documented for evaluation in accordance with the university’s academic standards.
AI Critique Activity — H1 · A2 · P · V · E
You may discuss concepts and possible approaches with classmates, but you may not share or review prompts, AI outputs, drafts, or completed analyses. Use the generative AI tool designated by the instructor (e.g., ChatGPT or Microsoft Copilot) to create the output you will critique, but complete the analysis independently.
Submit the prompt and output, identify and verify consequential factual or reasoning problems, and document any corrections. Conclude by explaining why your corrections are warranted and how they reflect the course concepts being assessed.
Faculty resources
The Faculty Guide in Canvas includes:
- complete framework definitions;
- a sample syllabus statement;
- assignment labels and Canvas-accessible code;
- guidance for selecting accountability evidence;
- Canvas implementation instructions; and
- frequently asked questions.
- Current Syllabus Guidance
- AI Guidelines for Instructors
- Generative AI in Your Classroom
- Academic Standards