University of Hamburg - to research, to teach, to educate and form, to homepage
Department of Mathematics

Photo: UHH/Denstorf

Guidelines for Students on the Use of Generative AI in Courses at the Department of Mathematics

  1. The purpose of your studies is to learn analytical, structured, and abstract thinking. This learning process depends on you thinking about mathematics yourself and solving exercises on your own, including mistakes and frustration.
  2. Sensible use of AI is possible, but it requires a fundamental understanding of the topics and concepts.
  3. Most people learn more and have more fun when they think about mathematics together. So if you get stuck on an exercise, ask your fellow students first before turning to a machine.
  4. Whenever you include AI-generated content in your submission, you must indicate this (just as with any other source).
  5. You remain responsible for the content and correctness of your work.
  6. Respect data protection when using AI. In particular, other people's data and documents should not be provided to an AI without their consent.
  7. The rules of the respective course are binding.

on 1) Learning to think

Studies (e.g. The Generative AI Learning Penalty: Evidence from Chinese Secondary Education) suggest that extensive use of AI tools to solve unfamiliar problems can help a person reach results faster, but that significantly less structural knowledge is acquired in the process; see, e.g., Figure 1 in https://cepr.org/voxeu/columns/generative-ai-learning-penalty-secondary-school.

The figures plot the distribution for students who eventually adopt generative AI before adoption (Pre AI), after adoption (Post AI) and students who have not adopted generative AI by June 2025 (Never AI).
© David Strömberg, Victor Lei, Yanhui Wu

Having an AI solve your exercises is like taking a forklift to the gym. It can move more weight, but nobody benefits.

When you read model solutions (whether written by a human or by an AI), you can easily (and wrongly) convince yourself that you could have solved everything on your own. Only by working on the problems yourself do you notice where the difficulties lie.

In surveys, graduates report that precisely the analytical, structured, and abstract thinking they developed through practice during their studies is also very important in their later professional lives.

on 2) Sensible use of AI

If, for example, you want to work on exam-style problems to prepare for an exam, you can ask an AI for suitable problems. You can then discuss your solutions with other students — and afterwards also have them checked by an AI, whose answers you should then carefully check yourself. However, this only helps you if you already understand the course well enough to work on the problems meaningfully, understand the feedback, and notice any errors made by the AI.

on 3) Learning together

A general piece of advice: Attend lectures and tutorials, and talk to your fellow students and instructors!

Talking about mathematics is also something you can (and must) learn. This skill not only helps with learning but is also very important in your later professional life (at university or elsewhere).

on 4) Cite your sources

This is a fundamental principle of academic work.

Search functions and spell checkers are unproblematic; if you adopt proof ideas or complete passages of proofs, you must state this.

Note that AI-generated texts often draw on sources without citing them. The responsibility for citing the correct sources remains with the user.

on 5) Check AI-generated content

This means that you must engage critically with AI-generated content. This concerns correctness, form, and relevance.

on 6) Respect data protection

For example, you should not upload lecture notes, exams, or solutions without asking their author. You should not upload documents containing the names of fellow students.

This document was prepared by the Working Group on AI & Teaching at the Department of Mathematics and will be updated as needed. Last updated: 8 October 2026.