Goal
Discuss how to distinguish productive GenAI use from unproductive dependence. The aim is to identify practical criteria: when does a student learn more, and when does the tool mainly replace the effort that learning requires?
Verification reminder. All mathematical claims, solutions, and code should be checked independently by a human.
Suggested task
- As a group, list 3 situations in which a student might use GenAI.
- For each situation, decide whether the use is mainly supportive, risky, or clearly harmful.
- Try to formulate one or two rules of thumb that could be given to students.
- Discuss how teachers might design assignments or guidance so that GenAI use supports learning rather than passivity.
Possible scenarios
- A student asks for five extra practice problems on a topic they are revising.
- A student pastes a homework problem and asks for the full solution immediately.
- A student asks for a different explanation after not understanding the textbook.
- A student asks the GenAI to rewrite their proof without checking whether the mathematics changed.
Example prompt
I am a university student learning real analysis. I do not want the full answer immediately. Ask me one question at a time so I can work through the problem myself. Only give a stronger hint if I ask for it explicitly.
Questions for the group
- What signs suggest that the student is still doing the intellectual work?
- What signs suggest that the student is outsourcing the thinking?
- Are there uses that are acceptable for practice but not for assessed work?
- What guidance could the department give that is realistic rather than purely prohibitive?
Minimum output expected from the group
- One acceptable use case
- One risky or unacceptable use case
- One concrete rule or guideline for students or staff
For a compact summary that a department could adapt, see Suggested departmental guidance.