MIT AI Academic Integrity (Ethics Standards)
At MIT, there is no single rule that makes AI use acceptable in every class or assignment. Check the syllabus and assignment instructions first, then ask your instructor if the limits are unclear. Use AI only as allowed, disclose it as required, and verify every part of your work. You remain responsible for what you submit.
Academic work depends on more than getting a correct answer. Your instructor may also be assessing how you reason, write, solve problems, or work on your own. AI can help with some tasks, but it can also blur who did the work and whether a course rule was followed.
I use a simple hypothetical to show why the details matter: a student may be allowed to use AI to brainstorm for one assignment but required to write another without AI help. The same tool, used by the same student, can be acceptable in one case and not in the other. The assignment rules decide.
MIT AI use and academic-integrity rules
Academic integrity means doing and presenting your work honestly, following the rules for the task, and giving credit where required. At MIT, AI use must fit the rules for the specific course and assignment. A general belief that “AI is allowed” or “AI is banned” cannot replace those instructions.
Start with MIT’s Academic Integrity Handbook and the rules your instructor gives you. Read the syllabus, assignment prompt, and any later clarifications. Course instructions may set specific limits on tools, tasks, collaboration, and disclosure.
“Independent work” means work you are expected to complete yourself, without help that the instructor has not allowed. If an assignment says to work independently, do not assume that AI help is permitted just because it is not a person. AI can still provide assistance that changes the nature of the work.
Check the rules before you enter a prompt or use an AI-generated answer. If the instructions are unclear, ask your instructor. Keep the answer, especially if it explains which tools and uses are permitted. A short written clarification can help you follow the same rule throughout the assignment.
A useful standard is:
Coursework requirements: permission, attribution, and disclosure
Permission answers whether you may use AI for a task. Attribution gives credit or identifies material from another source. Disclosure tells your instructor how you used a tool. These ideas are related, but they are not interchangeable: citing AI does not make a prohibited use acceptable.
Before starting, look for rules on:
- Whether AI tools may be used at all.
- Which tasks they may support, such as brainstorming, editing, coding, or translation.
- Whether AI may create text, code, calculations, or analysis.
- What disclosure or citation format the instructor requires.
- Whether you must submit drafts, prompts, or other records.
If permitted use is unclear, ask a specific question. For example: “May I use an AI tool to explain an error message, if I write and test the code myself?” This is more useful than asking only whether “AI is allowed,” because it identifies the task and the kind of help you want.
Permission is not the same as disclosure
Permission is the instructor’s approval to use AI for a defined purpose. Disclosure is your account of the use. If an instructor allows AI brainstorming but asks you to report it, both conditions apply. If AI is prohibited for a task, reporting that you used it does not turn the use into an approved one.
| Situation | What to do | Why it matters |
|---|---|---|
| The prompt allows AI for brainstorming but not drafting | Use it only to develop ideas; write the submission yourself | The permitted use has a clear limit |
| The syllabus is silent about AI | Ask the instructor before using it | Silence is not clear permission |
| AI edits grammar, and the instructor permits editing help | Follow the disclosure rule and review every change | Edits can alter meaning or claims |
| AI produces code for an individual coding task | Do not submit it unless that use is allowed | It may replace work you are meant to do |
| AI gives you a source or quotation | Check the original source yourself | AI can produce false references or details |
When disclosure is required, follow the instructor’s format. If no format is given, ask what details to include rather than inventing a citation style. A clear account may identify the tool, the task it helped with, and what you changed or checked, but your course’s instructions take priority.
Academic success: verification, privacy, and responsible tool use
Verification means checking AI-assisted material against reliable evidence and your own understanding. AI can make confident mistakes, including false citations, flawed calculations, and code that fails in a particular setting. You remain responsible for the submitted work, whether or not a tool helped produce it.
Check each part of the submission that matters:
- Facts and claims: Confirm them with trustworthy sources, not just another AI answer.
- Citations: Open the cited source and make sure it exists and supports the claim.
- Calculations: Rework important steps or test them with an approved method.
- Code: Read it, run relevant tests, and make sure you can explain its behavior.
- Writing: Check that the final version reflects your reasoning and does not claim more than you can support.
A tool’s answer is not proof that you understand the material. If you cannot explain a key claim, calculation, or code choice without relying on the tool, take time to learn it before submitting. This is especially important when an assignment tests your own understanding.
Protect information when using AI services. Do not enter sensitive personal information, confidential research, restricted course materials, or another person’s data unless you have permission to do so. If you are unsure whether information is restricted, leave it out and ask the instructor or the relevant office.
A practical record of AI use
Keeping a brief record can help you disclose use accurately and answer questions later. Follow any course directions about records or submission. Do not share private prompts or materials if doing so would violate a privacy rule.
You can note:
- The tool and date of use.
- The task it supported.
- The prompts or key instructions you gave it, if appropriate.
- Which output you used, changed, or rejected.
- How you checked facts, sources, calculations, or code.
This record is not a substitute for permission. It is a way to keep track of what happened so that your disclosure is accurate.
Scenarios: applying the rules to common assignments
These are examples, not reports of actual MIT cases. Each shows how the same AI tool can raise different questions depending on the assignment’s rules. The key step is to match the proposed use to the instructor’s directions, then ask before proceeding if the answer is uncertain.
Essay planning: Your instructor allows AI for brainstorming but says the submitted argument must be your own. You could ask for possible themes, then check the ideas and build your own argument. Do not copy generated paragraphs or citations unless the assignment specifically permits that use.
Programming assignment: The prompt says to complete the code independently. An AI tool that writes a function may do part of the task for you, even if you later edit the result. Ask whether help with explanations or debugging is permitted, and do not use generated code until you have a clear answer.
Language or grammar help: Your instructor allows proofreading tools. You may use an approved tool to flag grammar issues, but you still need to review its suggestions. If it rewrites a paragraph or changes your meaning, that may go beyond proofreading. Check whether the course treats rewriting differently from error correction.
Research sources: AI suggests a paper and gives a quotation. Search for the paper, read the relevant passage, and confirm the quotation and page details. If you cannot find the source, do not cite it. A citation that looks complete is not evidence that the source exists.
In each example, the safest next step is not to guess at a boundary. Identify the exact task, check the written rules, and ask for clarification when needed.
Research and career pathways: transferable AI ethics practices
Responsible AI use means being clear about what a tool contributed, checking its output, protecting information, and following the rules that apply. These habits matter in coursework and can also help you work responsibly in research or a future job. The exact rules may differ by setting, so always check the policy that governs the task.
Research work may involve private data, unpublished findings, or materials with limits on how they can be shared. Do not place those materials into an AI service unless the appropriate rules and permissions allow it. Ask a supervisor or instructor if you cannot tell whether the material is restricted.
For class assignments, use this sequence:
- Prepare: Read MIT’s academic-integrity guidance, your syllabus, and the assignment instructions.
- Clarify: Ask the instructor about any uncertain tool, task, or disclosure rule. Keep the response.
- Use within limits: Do only the permitted work with AI. Do not use a tool to bypass a requirement for independent effort.
- Verify: Check claims, sources, calculations, and code. Make sure you understand the final work.
- Disclose and retain records: Follow the course’s disclosure rules and keep relevant records if appropriate.
If you learn that your use may not have followed the rules, do not hide it or alter records. Review the course guidance and contact the instructor promptly to ask how to address the situation. The instructor can explain the next step for that course.
Conclusion: a reliable decision check
A sound decision about AI use begins with the assignment, not the tool. Confirm what is permitted, ask when the rule is unclear, and keep your work honest about who or what contributed. Then verify the final submission and disclose AI use in the way your instructor requires.
Before submitting, ask yourself:
- Did I check the rules for this assignment?
- Did I use AI only for a permitted purpose?
- Can I explain and support the work I am submitting?
- Did I verify sources, calculations, and code?
- Did I disclose AI use as required?
- Did I protect private or restricted information?
If you cannot answer yes to a relevant question, pause and resolve it before submitting.
Frequently asked questions
These answers give a quick starting point, but your instructor’s rules for the course and assignment control. When a written instruction does not answer your question, ask the instructor before using AI. A tool’s general terms or advice from another class cannot replace course-specific directions.
Does MIT ban students from using AI?
There is no single blanket rule that makes AI use permitted or prohibited for every MIT assignment. Check the course and assignment instructions.
Is AI allowed if the syllabus does not mention it?
Do not treat silence as permission. Ask the instructor whether your planned use is allowed.
Can I use AI if I cite it?
Only if that use is permitted. Citation or disclosure does not make prohibited assistance acceptable.
Do I have to disclose every AI use?
Follow the instructor’s disclosure rules. If they are unclear, ask what uses must be reported and how.
Can AI help me edit an essay?
That depends on the assignment rules. Some instructors may allow limited proofreading but not rewriting or drafting.
Can I submit AI-generated code?
Only when the instructor permits that use. If the task requires independent coding, generated code may replace work you are expected to do.
What if AI gives me a fake source?
Do not cite it. Find and read the original source, and confirm that it supports your claim.
Am I responsible for an AI mistake in submitted work?
Yes. You are responsible for checking the accuracy and support of what you submit, including AI-assisted material.
Should I enter course data into an AI tool?
Do not enter sensitive, confidential, or restricted information unless you are authorized to do so. Ask if you are unsure.
What should I do if I used AI before checking the rules?
Review the assignment instructions and contact your instructor promptly. Explain what tool you used and what it contributed, then ask how to proceed.
(This article was written by one of our staff writers, Alan Westbrook. Visit our Meet the Team page.)
