AI Policies for Student Assignments

This document makes recommendations for how faculty may communicate the appropriate use of AI by their students on course assessments and activities.  It does not cover any other use of AI by faculty.

There's no single "right" AI policy. Your approach should fit your course goals, your teaching style, and your students' needs. The most important things are to be clear, be consistent, explain your reasoning, and stay open to adjusting as you learn what works best.

Generative AI is a new and emerging technology.  Acceptable usage of AI is still an unsettled area of higher education pedagogy.  How and whether faculty choose to integrate AI usage in their courses is solely up to the discretion of the faculty member, which means that students are exposed to a large spectrum of subjective views on academic integrity in this area.  This differs from a more objective academic integrity area, such as plagiarism. Students want guidance on this issue. A thoughtful policy—even if it's strict—is better than leaving them uncertain about expectations.

Based on the wide spectrum of opinions that GCSU faculty have on what is considered acceptable and unacceptable use of AI in academic coursework, students should have the opportunity to clearly understand each faculty member’s expectations for acceptable and unacceptable AI use for a given course.  It is highly recommended that faculty provide clear guidelines for AI expectations for all course assessments and activities.  We recommend that faculty include a statement of use for every graded activity and assessment.  The statement of use should clearly state what aspects of AI usage are acceptable and/or unacceptable for the given activity or assessment. 

Key Questions to Ask Yourself

As you develop your statements, consider:

• What skills am I trying to help students develop?
• How does AI use support or interfere with those goals?
• Am I prepared to detect and address policy violations?
• Have I explained my reasoning in ways students will understand?
• Is my policy realistic and enforceable?
• Does my policy account for different student needs and backgrounds?
• Am I modeling good practices with AI in my own teaching?

A quality statement of use should address four core components: what students can use AI for, what is not allowed, which AI tools the policy covers, and what happens if the policy is violated. Together, these four pieces give students a complete picture of your expectations rather than a partial one — a statement that only says what's prohibited, for example, can still leave students unsure which tools count or what happens if they get it wrong. Each component is described below.

2.2.1 - What can students use AI for?

Think about which learning tasks benefit from AI assistance and which ones don't. Match your rules to your learning goals.

2.2.2 - What Is Not Allowed?

Be clear about prohibited uses. Students need to understand the boundaries.

Common prohibitions:

• Submitting AI text as your own work without attribution
• Using AI on assignments where it's not permitted
• Having AI complete entire assignments
• Using AI during exams (unless specified)
• Bypassing the learning process by over-relying on AI

2.2.3 - Which AI tools Does This Policy Cover? 

Be specific about what counts as AI in your policy. Students use many different tools, and they need to know what your policy covers.

Common AI Tools

• Text generators (ChatGPT, Claude, Copilot, Gemini)
• Grammar and writing assistants (Grammarly, QuillBot)
• Translation tools (DeepL, Google Translate)
• Tutoring platforms with AI features
• Coding assistants (GitHub Copilot, Tabnine)
• Search tools with AI summaries
• Image generators (DALL-E, Midjourney)

2.2.4 - What Happens If Students Break the Rules?

State consequences clearly so students understand the stakes:

• Following your institution's academic integrity policy
• Requiring assignment revision
• Grade penalties
• Educational conversations before penalties
• Different consequences for first-time versus repeat issues

The four components above should appear in every statement of use. The considerations below are optional refinements — use them only if your policy for a given assignment calls for staged permissions or documentation requirements.

When Can Students Use AI During an Assignment?

Some faculty allow AI at certain stages but not others. This helps students develop their own skills first.

Consider:

• Early stages only (brainstorming, outlining)
• Revision stages only (after creating a draft)
• Throughout the process with documentation
• Never for certain assignments
• Different rules for different assignment types

How Should Students Document AI Use?

If you allow AI, require students to show how they used it. This promotes transparency and helps you assess their learning.

Documentation methods:

• Include AI conversation transcripts in an appendix
• Add a statement describing what AI did and did not do
• Highlight or mark AI-generated content
• Explain how they modified AI output
• Submit both AI drafts and final versions
• Use a standard form or template you provide

AI Use Statement

For this assignment, AI tools may be used only during the brainstorming and outlining stage. You may use an AI tool (such as ChatGPT, Claude, or Gemini) to help you generate potential arguments, identify counterarguments, or organize your initial ideas before you begin drafting.

What is allowed:

• Using AI to brainstorm possible thesis positions or arguments
• Using AI to generate a list of potential counterarguments to address
• Using AI to help organize your outline structure

What is not allowed:

• Using AI to write any sentences that appear in your draft or final submission

• Using AI to generate analysis, evidence, or reasoning that you present as your own

• Using AI at any stage after you begin drafting, including for revision, editing, or proofreading

Acceptable and unacceptable tools:

• Acceptable: general-purpose AI chat tools (ChatGPT, Claude, Gemini, Copilot) used only for the brainstorming activities listed above
• Not acceptable for this assignment: AI writing assistants that generate or rewrite sentences (e.g., Grammarly's generative rewrite features, QuillBot), even during brainstorming
• Translation tools are permitted only for multilingual students accessing the assignment prompt itself, not for generating essay content

Documentation required: If you use AI during brainstorming, include a one-paragraph note at the end of your submission describing which tool you used and what you asked it to help with (e.g., "I used Claude to help me think of counterarguments to my thesis about renewable energy subsidies"). You do not need to submit a full transcript, but your note should be specific enough that I can see how AI fit into your process.

Consequences for policy violations: Any AI-generated writing found in a draft or final submission will be treated as a violation of this assignment's AI policy and referred to the university's academic integrity process. Before that referral, I will schedule a conversation with you to discuss the concern and give you the chance to explain your process. A first-time, undocumented use of AI for brainstorming (without evidence it affected your written analysis) will result in a request to add the required documentation rather than a formal violation.  Subsequent undocumented use of AI will be considered a violation of this assignment’s AI policy. 

Some faculty use a red/yellow/green color system as a quick visual shorthand for AI use expectations, often displayed at the top of an assignment sheet or in a syllabus. This is not a replacement for the statement of use outlined in 2.2 — it is an optional layer that can sit on top of one, giving students an at-a-glance signal before they read the full details.

Red and Green are more absolute and can often stand on their own.

• Red means AI use is not permitted at all for this assignment. Because this is an absolute prohibition, a Red label communicates most of what a student needs to know on its own. Faculty may still want a brief statement covering tool definitions (2.2.3) and consequences (2.2.4), but the "what's allowed" question is fully answered by the color itself.
• Green means AI use is allowed without significant restriction. Like Red, this is a largely absolute designation — the label itself tells students the main answer. Faculty should still be clear about documentation expectations if any apply, but Green does not require the same level of qualification that a partial-permission policy does.

Yellow requires a full statement of use.

Yellow signals that some AI use is acceptable and some is not — and that distinction cannot be captured by the color alone. A Yellow label without an accompanying statement leaves students guessing exactly where the line falls, which recreates the confusion this guide is meant to prevent. Any assignment labeled Yellow should be paired with a complete statement of use covering all four core components in 2.2: what's allowed, what's not, which tools count, and what happens if the policy is violated.

A note on accessibility. If you use this framework, always pair the color with the word itself (e.g., "Yellow" rather than a color alone), since red-green color blindness is common and a color-only label may not be visible to all students.

AI detection tools like Turnitin's AI Detector can be part of your assessment process, but they should never be the only evidence you use to determine whether a student used AI inappropriately. These tools are not perfectly accurate—they can flag student-written work as AI-generated (false positives) and miss actual AI use (false negatives). Think of AI detectors as a conversation starter, not a final verdict. If a detection tool flags an assignment, or if something about the work doesn't seem consistent with a student's typical writing, approach the student with curiosity rather than accusation.