AI and Academic Integrity: Keep Your Work Clear and Attributed

AI and academic integrity means using AI only within the rules for your course and assignment, acknowledging the assistance you received, and taking responsibility for the work you submit. A citation does not make prohibited assistance acceptable, and permission to use AI does not make its answers reliable.

Tools such as fenbs can keep the applicable rules and your study decisions outside a single chat. That record helps you remember what was allowed, what help you actually used, and which parts you still need to check. It supports honest work; it cannot grant permission or certify that an assignment meets your school’s requirements.

Which rules apply to this assignment?

Start with the institution’s academic integrity policy, your syllabus, and the assignment instructions. Read them together. A general statement allowing AI may have an exception for an exam, a first draft, or a particular stage of research. Keep the actual wording and its source available instead of relying on a summary from a classmate.

Cornell’s academic integrity guidance↗ (opens in a new tab) illustrates why assignment details matter: instructors can prohibit AI, permit it with attribution, or set different expectations for individual assignments. That is guidance for Cornell instructors, not a policy that applies to every US college. Your own instructor’s directions are the starting point for your work.

• Identify the permitted activity: brainstorming, explanation, editing, coding, or generating submission content.
• Check whether permission applies to this assignment and this stage of the work.
• Find the required disclosure or citation format and any instructions about retaining prompts.
• Ask the instructor before using a tool when the written rules leave a material question unanswered.

An answer from an AI assistant is not permission. If two instructions seem inconsistent, describe the specific conflict to your instructor rather than choosing the interpretation that permits the most help.

What assistance should you record?

Keep enough information to describe what happened accurately. You usually need a short, specific record more than a large folder of unexplained screenshots. Follow any course requirement to keep full prompts or transcripts, and preserve original drafts in the tools where you wrote them.

Record changes in direction too. If an assistant’s outline would replace the reasoning the assignment is meant to assess, stop and return to your own notes. Keeping work context in an AI brain can make that boundary easier to carry into the next session.

How do attribution and source citation differ?

Attribution explains the assistance used to produce the work. A source citation identifies the material supporting a claim or quotation. These can overlap, but they answer different questions. Saying that AI helped reorganize your outline does not tell a reader where your historical evidence came from.

The Modern Language Association’s guidance on describing AI use↗ (opens in a new tab) distinguishes acknowledgment from citation and recommends reading and citing original source material directly. Use the style and disclosure instructions your instructor requires. A polished reference entry is not a substitute for describing substantial assistance honestly.

For example, an allowed-use statement might explain that you asked an assistant to suggest objections to an argument, then chose and researched one yourself. Do not use that example as a universal required format. Explain the real activity, the part of the submission it affected, and the checks you performed.

Cornell’s AI attribution guidelines↗ (opens in a new tab) offer several documentation approaches. The practical lesson is to agree on expectations early, including assistance that changes your process without contributing a directly quoted sentence.

How can you check sources and preserve your own reasoning?

Treat an AI-provided reference as a lead until you locate and read it. Check that the work exists, that the author and publication details match, and that the cited passage supports the claim. If you cannot retrieve it, do not turn an assistant’s confident description into a source you pretend to have read.

1. Write your own provisional answer from the assigned material before requesting permitted feedback.
2. Separate suggestions from evidence. Mark an interesting AI-generated claim as unverified until you check it.
3. Open the original source, read the relevant passage in context, and record its location.
4. Explain the conclusion in your own words without looking at the assistant’s answer. Revisit anything you cannot explain.

Keep quotations distinct from paraphrases while taking notes. Otherwise a copied passage can lose its quotation marks between tools. A record of what you checked is useful, but the submitted citation still needs to follow the assignment’s style requirements. fenbs Decisions and rules can keep the applicable permission beside that verification work.

What should you check before submitting?

Review the actual submission, including appendices, code, images, and references. Ask whether every part follows the relevant permissions and whether the disclosure describes the assistance you used. Do not assume that changing the wording of generated content changes whether its use was allowed.

Then test your understanding. Can you explain why your method fits the question, defend your main evidence, and describe a limitation? If not, use the remaining time to learn that material. The discipline of reviewing assistance also carries into using AI at work, where another person may rely on the result.

If you discover a problem, follow the course’s process for clarification or correction. Keep your drafts and the relevant instructions available. Do not fabricate a history of your work or alter a record to make an unauthorized step disappear.

How to do this with fenbs

1. Create a free account, then add a project for the course or research assignment.
2. Record the applicable directions in Decisions and rules, with a link to the course source and the scope of any instructor clarification.
3. Add tasks for your draft, source verification, and required disclosure. State what is unfinished rather than labeling an unchecked answer complete.
4. Connect a supported assistant using Integrations. Ask it to read the project rules before doing any permitted work and save progress before stopping.

fenbs does not check for plagiarism, determine authorship, or replace your school’s submission system. Keep restricted course materials and private student information in approved institutional systems. For a single assignment with no ongoing handoff, a small local study log may be enough.

Where should you go for more specific help?

For adjacent tasks, see academic research tools, AI assistance for academic papers, and ChatGPT and academic integrity. Each still depends on the rules for the work you are actually doing.

Use the broader AI governance guides to think about accountability beyond one assignment. The useful habit is consistent: identify the rule, keep your contribution visible, verify the evidence, and describe the help honestly.