Teaching & Learning
- No university-wide rule for general coursework and assignments is defined in the provided sources
- A department-specific thesis rule allows AI use at levels A–D, requires critical evaluation of AI-generated elements, and prohibits full AI co-writing at level E
Students can use AI tools such as Grammarly or ChatGPT according to levels 1 to 4 of the AI Assessment Scale.
D AI TASK COMPLETION, HUMAN EVALUATION AI is used to generate certain elements of the thesis. The elements generated by AI must be critically evaluated and questioned.
E FULL AI AI is used as a kind of “co-pilot” with whom the entire work is written together. This stage is not allowed in the thesis.
- The provided sources do not set a single university-wide exam rule; they indicate that allowed AI use depends on the assessment format and permitted aids
- Instructor guidance material says instructors may prohibit generative AI in summative assessment, but also says a blanket ban would be legally unenforceable except in proctored exams
Prohibit its use as a tool/resource in summative assessment and take measures to enforce it.
Communicate to your students what tools or resources are not permissible for your assessment type and why you’ve decided not to allow generative AI tools like ChatGPT in your assessment practices.
We’d advise against a blanket ban on ChatGPT as it’d be legally unenforceable except in the context of proctored exams.
For examinations and credits, you must adhere to the permitted aids and mark them accordingly. So first find out whether the program is allowed as an aid for your current work
- AI use for non-graded learning support is described as permissible in the provided sources
- Departmental guidance explicitly allows brainstorming, high-level explanations, and generating examples for understanding, as long as text is not copied into the final document
2.2 Idea Exploration and Clarification
- Brainstorming research questions.
- Getting high-level explanations of concepts (without copying text).
- Generating examples to deepen your understanding (not to include verbatim in the final document).
- No university-wide rule for programming assignments is defined in the provided sources
- A department-specific thesis guideline permits programming support such as code completion and debugging only if the student understands, verifies, and takes responsibility for any submitted code, and prohibits producing code the student does not fully understand
2.3 Programming Support
- Code completion and debugging suggestions.
- Explanation of error messages.
Condition: You must understand, verify, and take responsibility for any code you submit.
- Producing code, models, or results you do not fully understand.
Research
- The sources indicate that AI may be used in research writing only in limited ways and with disclosure and human responsibility
- TUM materials say researchers should disclose AI tools and machine contributions, remain responsible for accuracy and conclusions, and a department guideline forbids using AI-generated text as a substitute for one's own academic writing
Researchers should disclose the tools and algorithms they used and clearly identify the contributions of machines and humans.
Researchers remain responsible for the accuracy of the data and the conclusions they draw
When it comes to phrasing texts, these programs can be very helpful. However, the question is whether they are allowed as aids for your current work and whether you adopt the AI-generated texts word-for-word or only use them as a basis for your own writing (AI-assisted vs. AI-generated).
Students are not allowed to use AI tools for any of the following:
3.1 Generating Substantial Parts of the Thesis, Seminar Paper, or Project Report
- Writing full paragraphs, sections, or the entire document.
- Using AI-generated text as a substitute for your own academic writing.
- TUM's AI strategy says large data use requires data-sharing guidelines and that data must be handled with the greatest care while preserving privacy and security
- The provided sources do not set detailed operational rules for AI-based research data analysis, but they do state governance expectations for data handling in AI work
Ein wichtiger Aspekt der KI-Entwicklung ist das Training von Modellen. Dafür werden große Datenmengen benötigt. Daher müssen Richtlinien für die gemeinsame Nutzung von Daten in Betracht gezogen werden, um Innovation und Zusammenarbeit zu erleichtern und sicherzustellen, dass die Daten mit der größtmöglichen Sorgfalt und unter Wahrung von Datenschutz und Sicherheit behandelt werden.
- TUM states that AI use in research must be transparent, scientifically sound, and monitored for harm and bias
- The AI strategy also says TUM will develop a code of conduct and ethical guidelines for third-party software, deep-learning solutions, public data, and AI use in publications and dissertations
Researchers should disclose the tools and algorithms they used and clearly identify the contributions of machines and humans.
Experts must ensure that their findings are scientifically sound and do no harm. For example, the risk of the AI being "biased" by the training data used must be kept to a minimum.
Finally, researchers, together with policymakers, civil society and business, should monitor the impact of AI and adapt methods and rules as necessary.
Darüber hinaus werden wir einen Verhaltenskodex und ethische Richtlinien für die Verwendung von Drittanbieter-Software, DL-Lösungen und öffentlich zugänglichen Daten (soziale Medien, Foren usw.) aufstellen. Auch ein Verhaltenskodex für die Nutzung von KI beim Verfassen von Veröffentlichungen, Dissertationen usw. soll entwickelt werden.
Academic Integrity
- Disclosure of AI use is required where covered by the provided policies
- Departmental guidance requires brief documentation of all AI use in a dedicated declaration, and the Educational Sciences thesis page requires the declaration of authorship to state which tools were used, at what level, for which thesis sections, and for what purpose; library guidance also says AI cannot be cited as a source of information
All AI use must be briefly documented in a dedicated section of the submitted document (e.g., "Declaration of AI Use"). It should include:
- Which tools were used (e.g., ChatGPT, Gemini, Copilot, Grammarly).
- For what purpose (e.g., proofreading, generating initial ideas, debugging code).
- How you ensured accuracy (e.g., manual verification, comparison with original sources).
To document AI use transparently, the declaration of authorship must state which AI tools have been used at which level (according to the AI Assessment Scale) for the introduction, theory, methods, results, and discussion sections including the purpose (the result produced after using AI, e.g., brainstorming ideas, formulating texts).
The number or combination of AI use levels A–D has no bearing on the final grade, but must be reported for documentation and transparency purposes.
You cannot use ChatGPT or other AI programs as sources of information.
- The source states such tools are not official plagiarism detectors, can be circumvented, and create data privacy concerns
- Instructor guidance says AI bans in assessment may be enforced through assessment design and declarations of original work, but it rejects reliance on AI-detection software
Introduce a Declaration of Original Work for students to confirm that they’ve worked on assessment task/s on their own without the help of third parties and without the use of unauthorized aids.
Choose a form of assessment that would make the use of ChatGPT impossible (e.g., oral exam, proctored written exam).
While there are software solutions available that claim to detect AI-generated content, these tools aren’t considered official plagiarism detectors, can be outsmarted, and present data privacy concerns.
Institutional & Administrative
- The provided sources show that TUM is planning and supporting AI use by lecturers and staff, but they do not provide a detailed operational policy on grading, feedback, or recommendation letters
- TUM's AI strategy says additional AI offerings will be created for lecturers and staff, and the assessment factsheet gives instructors guidance on how to design assessments in response to ChatGPT
Integration into the curriculum: AI will be successively integrated into academic programs and extracurricular offerings. Students will acquire essential skills in areas such as data analysis, machine learning, and algorithm development. Additional offerings will be created for lecturers and staff.
In this changing landscape, we’d like to offer guidance on approaches you can consider and how you can adapt your teaching and assessment in response to ChatGPT.
- TUM has an institutional AI data protection guideline and ties AI use to privacy and security requirements
- The sources also show institutionally supported tools and services, including a TUM-internal AI guideline, Microsoft M365/Copilot availability, and Grammarly availability for TUM members
TUM Data Protection Guideline for the use of AI
The use of AI-supported processes / algorithms and AI-supported tools is steadily increasing. Ki and data protection are not mutually exclusive. Nevertheless, some principles must be observed.
The Microsoft M365 Suite has been available to all employees and students since January. A service agreement was also concluded for this purpose.
Ein wichtiger Aspekt der KI-Entwicklung ist das Training von Modellen. Dafür werden große Datenmengen benötigt. Daher müssen Richtlinien für die gemeinsame Nutzung von Daten in Betracht gezogen werden, um Innovation und Zusammenarbeit zu erleichtern und sicherzustellen, dass die Daten mit der größtmöglichen Sorgfalt und unter Wahrung von Datenschutz und Sicherheit behandelt werden.
- TUM has a comprehensive institutional AI strategy through 2030
- It sets strategic priorities across curriculum integration, AI-supported learning environments, research infrastructure, operations and administration, and ethics and transparency, and it emphasizes responsible, transparent governance of AI technologies
The TUM AI strategy outlines the plans for implementing AI technologies in five key fields of action:
* Integration into the curriculum: AI will be successively integrated into academic programs and extracurricular offerings. Students will acquire essential skills in areas such as data analysis, machine learning, and algorithm development. Additional offerings will be created for lecturers and staff.
* AI-supported learning environments: TUM is developing intelligent assistance systems and AI-based learning and examination environments to promote personalized learning and adaptive learning environments.
* Research and development: AI will be indispensable in numerous disciplines to overcome complex challenges and gain new scientific insights. The necessary research infrastructure is, therefore, being continuously expanded.
* Operations and administration: AI-supported tools should make documentation, support, and general processes more efficient and effective and speed up day-to-day tasks.
* Ethics and transparency: Addressing ethical issues such as bias in AI algorithms, data protection, and the impact of AI on work processes plays a central role. The aim is to strengthen trust in AI systems and ensure transparency.
Da KI Wirtschaft und Gesellschaft verändert, ist es von entscheidender Bedeutung, eine wirksame Governance-Strategie für KI-Technologien zu entwickeln. Während sichergestellt wird, dass KI verantwortungsvoll und transparent entwickelt und eingesetzt wird, sollte die Governance-Politik gleichzeitig die weitere Forschung unterstützen und fördern.
DocuMark: Responsible AI Use for Academic Integrity
Knowing your institution's AI policy is step one. DocuMark helps enforce it fairly by empowering universities to manage AI-generated content, prevent cheating, and support student writing through responsible AI use.
Common Questions About TU Munich's AI Policies
TU Munich has defined AI policies in 12 of 12 categories, with an overall coverage score of 100%.
Disclosure of AI use is required where covered by the provided policies. Departmental guidance requires brief documentation of all AI use in a dedicated declaration, and the Educational Sciences thesis page requires the declaration of authorship to state which tools were used, at what level, for which thesis sections, and for what purpose; library guidance also says AI cannot be cited as a source of information.
Instructor guidance says AI bans in assessment may be enforced through assessment design and declarations of original work, but it rejects reliance on AI-detection software. The source states such tools are not official plagiarism detectors, can be circumvented, and create data privacy concerns.
TUM has an institutional AI data protection guideline and ties AI use to privacy and security requirements. The sources also show institutionally supported tools and services, including a TUM-internal AI guideline, Microsoft M365/Copilot availability, and Grammarly availability for TUM members.