Rules & Guidelines

• Rules on the Use of AI in Academic Research
• Rules on Academic Collaborations
• Rules on Seminars (讨论班)
• Rules on Internships

Rules on the Use of AI in Academic Research

Prof. ZHANG Zaikun, March 25, 2026

0. You are encouraged to use AI, subject to academic integrity requirements. AI may assist your work, but it should not replace your understanding or responsibility.
1. You must understand every detail of the substantive mathematical content in your work, including but not limited to definitions, assumptions, algorithms, propositions, proofs, and derivations. You must be able to sketch and explain the reasoning, and reproduce the significant results and their proofs.
2. You must be able to explain every sentence in your writing: what it means, why it is there, and how it relates to the rest of the work.
3. You must be able to explain every substantively significant word or expression in your writing: what it means, why it is there, and why it is chosen instead of alternatives. (Before the AI era, this applied to every word and expression, not only the significant ones.)
4. You must be able to explain every substantive component (functions, subroutines, classes, files, etc.) of your code: what it does, how it works, and its role in the program. (Before the AI era, this applied to every line of code, not only the substantive components.)
5. Rules 2 and 3 also apply to comments and documentation of your code.
6. You must verify every factual claim, calculation, reference, and citation in your work. Any error or misinformation is your responsibility, not AI’s.
7. You must disclose material use of AI when required or academically relevant. When asked, you must be able to explain what tools were used, for what purposes, and how their outputs were checked.
8. Using AI does not reduce your responsibility for your work. AI-assisted material that violates these rules is unacceptable, and I will not endorse it.
9. If a thesis violates these rules, I will not approve its submission, or I will inform the committee about the violation and let it decide.
10. These rules are necessary conditions, not sufficient ones. You do not earn academic credit simply by understanding or reproducing a piece of work. The work must be your own, as should be common sense.

References (added on October 7, 2026)

[1] Leiden Declaration on Artificial Intelligence and Mathematics

[2] ICIAM Statement on Mathematics and Artificial Intelligence — Discovery, Understanding and Trust

[3] A Severe Misalignment of AI in Mathematics

[4] SIAM Publications — Editorial Policy on Artificial Intelligence

Rules on Academic Collaborations

Prof. ZHANG Zaikun, October 8, 2026

Collaborations are encouraged, subject to the following rules. In what follows, “the group” means my research group (L’Équipe Z), whose members include students (undergraduate and graduate), postdocs, research staff, engineers, etc.

1. Ideas arising from discussions within the group should be treated as intellectual property of the group unless otherwise agreed.
2. Internal discussions, unpublished ideas, or unpublished material of the group must not be disclosed outside the group without my permission. Ordinary discussions or presentations of ongoing research are encouraged, but I must be consulted in advance.
3. Collaborations outside the group must be discussed with me and agreed upon beforehand. Such collaborations are indeed encouraged, but I must be consulted in advance.
4. I will not be a coauthor of any work unless I consider myself to have made a substantive contribution. However, even when I am not a coauthor, members of the group are supported by my research funding, and their work should therefore acknowledge such support appropriately.
5. By default, papers published by the group list coauthors in alphabetical order.

Rules on Seminars (讨论班)

L’Équipe Z, January 21, 2026

1. Each speaker has one hour by default (presentation + Q&A). If more time is needed, the speakers should compromise among themselves in advance. The seminar will not extend beyond 18:00.
2. Prepare well-structured slides with solid supporting materials.
3. The presentation must begin with a review of the TODO list from the last presentation. It should contain the progress of each item and plans for unfinished ones.
4. Prepare the presentation meticulously! The presentation will be terminated in the following cases.
   • Three basic errors are found in the materials, e.g., typos, grammatical errors, obvious mathematical/logical mistakes.
   • The speaker fails to answer three basic questions, e.g., on undergraduate-level mathematics, motivation or implication of the work, basic tools used. Let AI assist you with the preparation! Include this rule on your first slide.
5. Within 24 hours after the presentation, send a TODO list and the revised slides to all group members.
6. To reschedule, inform the coordinator at least one week in advance. Avoid changes unless necessary.
7. Each group member is expected to ask at least one question during each presentation.

Rules on Internships

Prof. ZHANG Zaikun, May 9, 2026

Most of my graduate students are recruited only after expressing an intention to pursue academic research. Nevertheless, some may later change their plans, seek careers in industry instead, and therefore wish to undertake internships. Internships may be permitted subject to the following requirements.

1. Before taking an internship, the student must obtain formal written permission from the relevant institution (department, faculty, university, etc.) and government authority, as applicable.
2. During an internship, the student must not sacrifice the quality or quantity of the research required for their graduate study.
3. Academic requirements for graduation will not be lowered due to an internship.