Universities / Cornell University
Cornell AI statement: documenting AI use
At Cornell, each course sets its own AI rules, often shown with course policy icons such as UA (use with attribution) or AI-FREE. Cornell's Code of Academic Integrity, in force from August 24, 2026, counts any AI help your instructor hasn't explicitly allowed as unauthorized assistance. When a course allows use with attribution, the Center for Teaching Innovation suggests a paragraph or appendix: which tool and version, how and where you used it, your prompts, and how you checked the results. Instructors may instead ask for a traditional citation, a reflection, or the full transcript.
Independent guide, not an official Cornell University page: we aren't affiliated with or endorsed by Cornell University. Summarised from its official guidance, last checked 20 November 2026. Rules can change, so your unit outline or assignment brief wins if it says something different · See the full rules
Write yours
Cornell University doesn't publish a fixed template, so this uses a general format that covers what it asks for.
Cornell University's own wording
For assignments and projects, document how you used AI in a paragraph/appendix as transparently as possible, explaining: AI tool(s) used:Include the date, model/version used How did you use AI? (e.g., to generate ideas, turns of phrase, elements of text, long stretches of text, lines of argument, pieces of evidence, maps of conceptual territory, illustrations of key concepts, programming codes, etc.) Where in this assignment did you use AI? Describe the prompt(s)/parameter(s)/or preferences(s) you used to get results. How did you verify the information? (How did you confirm or challenge the results?)
Quoted from Cornell University's guidance.
Cornell University's own examples
Quoted word for word. Pick the one closest to what you did.
Traditional reference: APA-style in-text quote and reference entry (CTI sample)
When prompted with, “Is it ethical to use generative AI without proper attribution?” ChatGPT indicated, “Using generative AI without proper attribution can be considered ethically problematic, as it raises issues related to intellectual property, transparency, and honesty” (OpenAI, 2023). OpenAI. (2023) ChatGPT (Aug 10 GPT3.5 version) [Large language model]. https://chat.openai.com
Explanatory statement: how CTI suggests listing the tool, date and query
Chat-GPT-3. (YYYY, Month DD of query). 'Text of your query.' Generated using OpenAI. https://chat.openai.com/
Reflection statement: sample question prompts
For assignments and projects, document your AI use in a paragraph/appendix as transparently as possible, explaining: Why did you use these tools? (e.g., to save time, to surmount writer’s block, to stimulate thinking, to handle mounting stress, to clarify prose, to translate text, to experiment for fun, etc.) How did you use them? (e.g., to generate ideas, turns of phrase, elements of text, long stretches of text, lines of argument, pieces of evidence, maps of conceptual territory, illustrations of key concepts, programming codes, etc.) How did you verify the information? (How did you confirm or challenge the results?) What other (course) resources did you try before (or in addition to) using AI? Will you use this tool again in the future? Were you satisfied with the results? Did the tool successfully address the scope of the assignment?
Source: Cornell University's guidance.
Worked example
Illustrative example: AI use statement (appendix). AI tool used: ChatGPT (OpenAI, GPT-4o), March 3, 2026. How I used AI: to generate ideas for my research question and to suggest counterarguments. Where: Section 1 (topic choice) and the counterargument paragraph in Section 3. Prompts: "Suggest five research questions about food insecurity among college students" and "What are the strongest objections to universal free school meals?" How I verified it: I checked each counterargument against two peer-reviewed articles from the library and dropped one claim I couldn't confirm.
The rules at a glance
- When you need one
- When your course permits AI use with attribution (the UA icon), follow your instructor's format. The Code of Academic Integrity also says all outside assistance should be acknowledged.
- Framework
- Cornell's AI course policy icons: ANY-AI (any tool, any use), AT (approved tools only), UA (use with attribution), PP (privacy protecting, no proprietary materials), AS (assignment-specific), ER (explore and review only) and AI-FREE. Icons can be combined, for example AS+AT+UA. For attribution, CTI suggests four formats instructors can mix: a traditional reference, an explanatory statement, a reflection statement, or an extended / full transcript.
- Where it goes
- A paragraph or appendix in the assignment.
- Prompts and outputs
- It depends on the format your instructor picks. The explanatory statement asks you to describe your prompts, parameters or preferences. The full-transcript format asks for the prompts plus the entire exchange (screenshots or copied text) with the relevant parts highlighted, and a flowchart for heavy AI use in images, video or design work. CTI also suggests instructors may ask for chat logs with the assignment.
- Referencing
- Treat exact or modified AI text as a source: quotation marks, in-text citations and references in your discipline's style. CTI points to APA, MLA and Chicago, but notes these only suggest basic citations, so your instructor may ask for more detail.
- Must include
- AI tool(s) used, with the date and model/version
- how you used AI
- where in the assignment you used it
- the prompts, parameters or preferences you used
- how you verified the information
- Also note
- Follow your course's AI policy; it can differ between courses and between assignments
- From August 24, 2026, AI help not explicitly permitted by your instructor counts as unauthorized assistance under the Code of Academic Integrity
- You're responsible for the accuracy and existence of every source you cite, including fake citations generated by AI
- Don't upload sensitive personal information, confidential data (such as experimental results or unpublished articles) or non-public information into AI tools; use campus-approved tools for sensitive material
- Uploading copyrighted material into AI tools without permission might violate copyright law; under the PP icon, don't upload copyrighted or proprietary class materials
Questions about Cornell University's rules
- What do Cornell's AI icons like UA, AS or AI-FREE mean?
- They're markers instructors put on syllabi and assignments. UA means use with attribution, AT means instructor-approved tools only, AS means the rules change by assignment, and AI-FREE means no AI for any use unless stated. Icons can be combined, so AS+AT+UA means AI is allowed on certain assignments, with approved tools, if you attribute it. The icons go with a written course policy, so read that too. Source
- What happens if I use AI when my Cornell course doesn't allow it?
- From August 24, 2026, the Code of Academic Integrity treats AI help your instructor didn't explicitly permit as unauthorized assistance. For a minor case, such as unauthorized AI use on an assignment, your instructor may offer the Accepting Responsibility route, which means an in-person workshop and a penalty limited to that coursework. Otherwise there's a primary hearing, and if the evidence is clear and convincing the instructor can impose any grade penalty, up to failing the course. Hearing Boards can also recommend probation, suspension or a note on your transcript. Source
- Can a Cornell instructor accuse me based on an AI detector score?
- CTI doesn't recommend AI detectors for integrity cases, and says scores like Pangram or GPTZero shouldn't be the main evidence. Instructors are told to look for objective signs instead: citations that don't exist, work you can't explain when asked, or work that doesn't match your drafts. If your assignment asked for AI chat logs, those can be used as evidence too. Source
- Do I have to include my prompts in a Cornell AI statement?
- It depends on the format your instructor asks for. A traditional reference just quotes and cites the AI output. The explanatory statement asks you to describe your prompts, and the full-transcript format asks for the whole exchange, such as screenshots or copied text. A reflection statement focuses on why you used AI and how you checked it. Source
- Can I paste my research data or class readings into ChatGPT?
- CTI says to avoid putting sensitive personal information, confidential data like experimental results or unpublished articles, and non-public information into AI tools. Uploading copyrighted material without permission might also break copyright law. For sensitive material, use Cornell's campus-approved AI tools, which don't feed your material back into model training. Source
More help
Sources
- https://teaching.cornell.edu/generative-artificial-intelligence/ai-attribution-guidelines
- https://teaching.cornell.edu/generative-artificial-intelligence/ai-course-policy-icons
- https://teaching.cornell.edu/generative-artificial-intelligence/ai-academic-integrity
- https://deanoffaculty.cornell.edu/faculty-and-academic-affairs/academic-integrity/upcoming-code-of-academic-integrity/
Your unit outline or assignment brief wins if it says something different. Spotted a change? Tell us.