Assignment Design, Academic Integrity and Student AI Use

Generative AI has created new questions about authorship, academic integrity, and how faculty gather meaningful evidence of student learning. While no assignment can be made completely “AI-proof,” thoughtful design can make the learning process more visible and clarify what students are expected to do independently.

The goal is not simply to prevent inappropriate AI use. It is to create assignments that help students understand the purpose of their work, practice essential skills, and demonstrate their own thinking.

Begin with the purpose of the assignment

Before revising an assignment, identify the learning students are expected to demonstrate. Consider:

• What should students know or be able to do independently?
• Which decisions or thinking processes are most important?
• Could AI complete the central work of the assignment?
• Where might AI appropriately support the process?
• What evidence would show that meaningful learning occurred?

If AI can easily produce the entire final product, consider shifting greater attention to the decisions, reasoning, application, and revision that lead to that product.

Design assignments that make learning visible

Thoughtful assignment design can provide stronger evidence of student learning while also supporting student success.

Clarify the purpose

Explain why students are completing the assignment, what they are expected to learn, and how the work connects to the course.

Students are more likely to make responsible choices when they understand that the process itself is part of the learning.

Build in checkpoints

Consider dividing larger assignments into smaller stages, such as:

• Topic proposals
• Research questions
• Annotated sources
• Outlines
• Drafts
• Peer feedback
• Progress updates
• Individual conferences

Checkpoints allow faculty to provide guidance and better understand how students developed their work.

Ask students to explain their decisions

Include opportunities for students to describe their reasoning. They might explain:

• Why they selected a topic, method, source, or example
• How they applied course concepts
• What alternatives did they consider
• How they responded to feedback
• What they revised and why
• What they found challenging

These explanations can be included in a short reflection, a presentation, a process note, or a conversation.

Connect work to the course

Assignments are more meaningful when students must draw on specific course experiences. Ask students to incorporate:

• Class discussions
• Course readings
• Local or current examples
• Labs, fieldwork, or demonstrations
• Discipline-specific methods
• Data provided in the course
• Personal observations or professional contexts

Course-specific connections can encourage deeper application rather than generic responses.

Assess both product and process

When appropriate, evaluate more than the final submission. A rubric might include criteria related to:

• Development of ideas
• Use of evidence
• Application of course concepts
• Revision
• Decision-making
• Reflection
• Appropriate documentation of AI use

This approach reinforces that learning includes how students arrive at the final product.

Include assignment-level AI expectations

A general syllabus statement provides an important foundation, but expectations may differ from one assignment to another. For each major assignment, clearly identify:

• Whether AI use is permitted
• Which types of AI use are acceptable
• Which uses are not permitted
• Which parts must be completed independently
• How students should acknowledge AI use
• What students remain responsible for verifying
• What students should do when they are unsure

An example of a  statement might say:

AI use for this assignment: You may use generative AI to brainstorm possible topics and receive feedback on the organization of your draft. You may not use AI to write the final response, generate sources, or complete the analysis. You are responsible for verifying all information and briefly describing how you used AI when you submit your work.

Avoid relying on broad phrases such as “AI is allowed” or “AI is prohibited” without explaining what those expectations mean in practice.

Ask students to acknowledge AI use

When students are permitted or required to use AI, ask them to describe how the tool contributed to their work. A short acknowledgment might include:

• The tool used
• The purpose for using it
• The prompts or types of questions submitted
• What information or suggestions were used
• What was revised, rejected, or verified
• How the student remained responsible for the final work

For example:

I used Microsoft Copilot to brainstorm possible topics and generate questions to guide my research. I selected and revised two of the suggested questions. I did not use AI to write the final paper or locate sources.

Acknowledgment helps make the process visible and reinforces transparency without assuming that every use of AI represents misconduct.

When AI is part of the learning activity

AI can also be intentionally integrated into assignments. Students might be asked to:

• Critique an AI-generated response
• Correct inaccurate information
• Identify bias or missing perspectives
• Compare AI output with scholarly sources
• Revise generated content using disciplinary criteria
• Test how different prompts affect the response
• Reflect on when AI was helpful or unhelpful

In these activities, the learning comes from the student’s evaluation, judgment, and revision, not from the generated response alone.

Responding to possible inappropriate AI use

Suspected AI use can be difficult to evaluate. AI detection tools and writing patterns should not be treated as definitive evidence on their own.  Begin by reviewing the assignment expectations, the student's other work, earlier drafts or related assignments, sources and citations, and, when possible, the document history.

If you suspect a student has improperly used AI, begin with a conversation rather than an accusation. You might ask questions such as:

• “Can you walk me through how you approached this assignment?”
• “What tools or resources did you use while developing your work?”
• “Can you explain how you arrived at this conclusion?”
• “How did you select and verify these sources?”
• “What did your drafting and revision process look like?”

These questions can help clarify the student’s process and provide information that may not be visible in the final product.

Focus on evidence and university procedures

If concerns remain after speaking with the student, focus on specific evidence connected to the assignment and follow current CMU academic integrity procedures. Avoid deciding based only on AI detection scores, a sudden change in writing style, unusually polished language, or assumptions about what a student is capable of producing. These factors may raise questions, but they do not, by themselves, establish inappropriate AI use.

A fair response should consider the clarity of the original expectations, the available evidence, and the student’s explanation of their process.

Use AI detection tools with caution

AI detection tools may produce false positives and false negatives. Their results can be affected by writing style, language background, revision, and the type of content being analyzed.

Detection results should not be treated as proof. At most, they may serve as one piece of information that prompts further review and conversation.

Faculty should avoid submitting student work to third-party detection tools when doing so could create privacy, copyright, or data security concerns.

Support students before problems occur

Many students are still learning how AI expectations vary across courses and assignments. Faculty can reduce confusion by:

• Discussing AI expectations early
• Revisiting expectations before major assignments
• Providing specific examples of permitted and prohibited use
• Explaining why independent work matters
• Showing students how to acknowledge AI use
• Inviting questions before students submit work
• Providing alternatives when access or privacy is a concern

Clear guidance supports academic integrity more effectively than relying only on consequences after a problem occurs.

Need help?

The Office of Curriculum and Instructional Support can help CMU faculty:

• Review assignments for an AI-enabled learning environment
• Develop clear assignment-level AI expectations
• Create student AI acknowledgment language
• Add checkpoints, reflection, or process documentation
• Design activities that build AI literacy
• Consider appropriate responses to suspected AI misuse
• Align assignment design with TILT or backward design