How College Professors Can Use ChatGPT: Lectures to Feedback

Quick answer: College professors can use ChatGPT for the first draft of almost any routine academic task: lecture outlines, examples and analogies, quiz and exam questions, rubrics, reading summaries, abstracts and feedback wording. Give it the course, the level and the goal, check every result, and keep the decisions yourself. Do not paste student work or unpublished research into a public AI tool unless your institution says you may.
A lecturer's week is full of writing that is necessary but not the real work: slides, questions, emails, rubrics, comments. ChatGPT is good at the draft stage of these. It is weaker where academic work matters most, which is facts, sources and judgement. This guide shows where it helps, where it does not, and the rules to follow.
If you teach at school level, start with our guide on how teachers can use ChatGPT. This one is for college and university teaching and research.
Check your institution's rules first
Universities are still writing their AI rules, and they differ a lot between institutions, departments and even courses. Before you use any of the ideas below, find out three things:
- Student work. Is uploading essays, answers or projects to an AI tool allowed? Many institutions say no, because of student privacy.
- Research. Do your university, your journal or your funder set limits on AI in manuscripts, peer review or grant proposals? Several do, and some forbid putting confidential manuscripts into AI tools.
- Teaching. Does your department say how students may use AI in coursework, and how you must say so in a syllabus?
If you cannot find a written rule, ask your head of department or the office that handles academic integrity. Where you are unsure, the safe choice is to use ChatGPT only with material you wrote yourself or that is already public.
Lectures: planning and explaining
Build a lecture outline
You are an experienced [economics] lecturer. I teach [second-year undergraduates] and the class is [50 minutes]. Create an outline for a lecture on [price elasticity of demand]. Include: 3 learning outcomes, a 5-minute opening question, 3 main points with one real-world example each, one in-class activity for pairs, and a 3-question check at the end. Show it as a table with timings. Mark anything you are unsure about so I can check it.
You still decide what is true and what fits your syllabus. The outline simply saves the blank-page problem.
Explain a hard idea at different levels
Ask for the same concept three ways: for a first-year, for a final-year student, and for a student from another discipline. Pick the version that works, then add your own example.
Explain [Bayes' theorem] in three ways: (1) for a first-year student with no maths background, using an everyday example; (2) for a final-year student, with the formula and a worked example; (3) for a student from [a biology] background, using an example from their field. Keep each under 150 words.
Make examples, cases and analogies
New examples keep a repeated course fresh. Ask for five short examples from different industries or countries, then check that the facts in each one are real before you use it in class.
Questions for quizzes and exams
ChatGPT can draft multiple-choice questions, short-answer questions and marking points. Read every question yourself. Check that the right answer is right, and that the wrong options are wrong in a believable way.
Write [8] multiple-choice questions on [the causes of the 2008 financial crisis] for [third-year undergraduates]. For each: 4 options, the correct answer, and a one-line reason why each wrong option is wrong. Make two of the questions harder, requiring application and not recall.
Research: use it for support, not for facts
This is where professors need the most care. ChatGPT writes fluent text whether or not it is correct, and it can produce references that look real but do not exist. Treat it as a thinking partner and a writing aid, not a source.
Good uses
- Mapping a topic. Ask for the main debates, key terms and search phrases in an unfamiliar area, and then search for the papers yourself in your library or a database.
- Explaining a method. Ask it to explain a statistical test or a coding approach in plain words, then confirm against a textbook or the original paper.
- Drafting from your own notes. Paste your own rough notes and ask for a structured abstract, an outline or a clearer paragraph.
- Editing your writing. Ask it to shorten, clarify or fix the flow of text you wrote.
Here are my rough notes for a literature review section on [teacher motivation in higher education]: [paste your own notes]. Group them into 3 to 4 themes, suggest a logical order, and list what seems to be missing. Do not add any references or facts that are not in my notes.
Poor uses
- Asking for a list of references and citing them without opening each one.
- Using it to summarise a paper you have not read yourself.
- Putting unpublished data, a manuscript under review or a grant proposal into a public tool.
For finding and reading sources, a tool that shows where each answer comes from is easier to check. Our Perplexity AI tutorial shows how to research with citations, and Gemini and NotebookLM for beginners shows how to ask questions of documents you upload yourself, such as published papers and your own notes.
Feedback and assessment
Write a rubric
Create a marking rubric for [a 2,000-word essay] in [a second-year sociology course], out of [100] marks. Use a table with 4 criteria, 4 performance levels and a short description for each level. Use language students can understand. The criteria are: [argument, use of evidence, structure, referencing].
Share the rubric with students before the task. Then adjust the wording so it matches what you actually taught.
Draft feedback comments without student data
You can build a bank of comments without ever pasting a student's work. Describe a common weakness and ask for feedback you can adapt.
Write 5 short, constructive feedback comments for an undergraduate essay where the main problem is [a clear argument that is not supported by evidence]. Each comment should say what is good, what to fix, and one next step. Keep each under 60 words. Do not use a student's name.
Save the ones that sound like you. Adapt each one to the work in front of you when you mark.
What not to hand over
Do not let ChatGPT assign marks, and do not paste identifiable student work into it. A grade is an academic judgement, and it is yours to make and defend.
Communication and admin
- Emails. Draft replies to common student questions, extension requests and reference requests, then edit so they sound like you.
- Course documents. Turn your bullet points into a clear syllabus section, assignment brief or handout.
- Meetings and reports. Turn your own notes into minutes or a short summary.
Write a polite email, under 120 words, replying to a student who asks for an extension on [a coursework deadline]. Follow this policy: [paste your own policy]. Ask them to send [documentation]. Do not promise the extension.
How to write better prompts
- Role. Say who it should be, such as "an experienced lecturer in [subject]".
- Context. Give the level, the course and what students already know.
- Task. Say exactly what to produce and what it must include.
- Format. Ask for a table, a numbered list or a one-page handout.
- Limits. Add rules such as "do not add references" or "mark what you are unsure about".
If the first answer is not right, reply in the same chat and say what to change: "Make it shorter," or "Use a different example." Most good results come after one or two follow-ups. For more on this, see our prompt engineering templates.
What ChatGPT cannot do for a professor
- It does not know your students, your syllabus or your department's rules unless you tell it.
- It cannot be trusted with facts, numbers or references without checking.
- It cannot make an academic judgement, such as whether work deserves a pass.
- It does not replace your subject knowledge, which is how you notice when an answer is wrong.
How to start this week
Pick the task that takes the most time and has the least risk, such as a rubric or a set of quiz questions. Try one prompt from this page, check the result, and keep the version that worked. Add research uses only after you have read your institution's and publisher's rules.
Learn these tools step by step
AivaanTech's AI Tool Mastery course teaches ChatGPT, Claude, Gemini, Perplexity, NotebookLM and other tools through hands-on tasks. It is built for beginners, needs no coding, and ends in a capstone project and a certificate that anyone can verify online. Lessons are self-paced, and there are live classes and a Saturday Q&A, so your questions get answered.
See how the course fits a teacher's week on our page for AI for teachers.
Frequently asked questions
How can college professors use ChatGPT?
Can professors use ChatGPT for research?
Is it allowed to upload student work to ChatGPT?
Can ChatGPT grade student assignments?
How do I stop ChatGPT making up references?
Do I need a paid ChatGPT plan to try these ideas?
Ready to go deeper?
Explore the full 30-day AI Tool Mastery certification — prompt engineering, 11+ tools and a real capstone project.
