Responsible use of AI — guide for participants#
Generative AI is a tool you direct. It is not a substitute for your own thinking, your own hands, or the conversations you have with the people around you. Use it to go further than you could alone, never to skip the work that makes you capable of going further. Below you will find where AI is required, permitted with disclosure, or not permitted.
If you have further thoughts or different perspectives, please reach out to us, we are humans.
Dos#
-
Do the thinking first, then use AI to extend it. Sketch your own concept, structure, code path or recipe. We have a huge knowledge database at the Tutorials section of our open archive. Use AI to test alternatives, explain what you’re stuck on, or stress-test a draft, AI should not produce the first version for you.
-
Never use AI for inspiration references. Those define who you are and what you like. Mention the use of AI in every assignment. Add one short: AI use line per assignment page: which tool, what you asked for, what you kept, what you changed and why. Disclosed use is normal practice; undisclosed use is misrepresentation of authorship.
Always verify sources. AI tends to invent citations and fabricate false information, it can be misleading. Always double check every fact, citation, formula, material spec and machine setting against the original document, the datasheet, or your own physical test. -
In a hands-on programme, the fabrication, measurements, failures and fixes are the learning and should be documented. AI-generated processes, visuals or reflections cannot stand in for a process you didn’t actually run. Keep the making and the judgement in your own hands.
-
Use AI to prepare better questions, then take them to your instructor, the lab, your peers, the node. Those conversations build judgement that no AI model can give you.
Try it yourself first. Do not underestimate yourself, try to do any task yourself first, based on tutorials, search engine results. Then if you struggle, ask a precise question to an AI to solve this precise issue
Don’ts#
-
Don’t submit AI output as your own work, no pasted text, code, documentation or reflection presented as authored by you without disclosure and proper attribution.
-
Don’t ask AI to write up work you didn’t do.
-
Don’t outsource the struggle.
-
When you’re stuck, the first attempt is yours. Reaching for a chatbot at the first hint of difficulty feels like progress and costs you mastery and confidence.
-
Don’t put confidential material into public AI tools. Never paste unpublished material, other people’s personal data, or anything covered by confidentiality into a tool you don’t control.
-
Don’t cite what you haven’t read, and never present AI as an author or source. AI is a tool, not a co-author or a reference. Read the source, then cite it. Honest disclosure will never count against you; hiding it will.
Grounded in MIT’s Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training (final report, August 2026) — augmentation over automation, “no one size fits all”, productive struggle as the point of the education, instructor and student disclosure, and “AI should never be listed as a co-author.”
Sources: MIT president’s letter and the Committee report.