Yes – you may use AI tools for learning, as long as you use them thoughtfully, responsibly, and in compliance with legal requirements.

Artificial intelligence can support independent learning in many ways: it can provide prompts for reflection, explain complex topics in simple terms, or ask targeted follow-up questions. When used effectively, it not only enhances understanding but also promotes self-reflection. However, it is important not to accept the results of AI tools uncritically, but to use them as a complementary aid.

AI systems like ChatGPT rely on large datasets that may contain inaccurate, outdated, or biased information. Therefore, AI cannot replace academic sources or discussions with instructors or fellow students.

You should also consider copyright and data protection aspects: do not use sensitive data, and ensure that you respect the appropriate rights when reusing texts or sources (for more information, see the next question or the university’s AI guidelines for teaching).

Examples:

• AI as a Socratic tutor: Ask the AI to pose open-ended questions on a topic to test your understanding, encourage you to draw your own conclusions, or deepen your knowledge.
• Reflective learning journal: Have the AI suggest questions such as “What surprised you today?” or “What did you find difficult and why?” This can help you recognize and document your learning progress.
• Argumentation training: Ask the AI to deliberately generate flawed arguments and practice refuting them.
• Exam preparation: Use AI to create short quizzes based on your notes or learning objectives.

Course materials such as lecture notes or recordings may only be uploaded and processed with the instructor’s consent—because many of these materials are protected by copyright.

Lecture notes, slides, recordings, or past exams are generally the intellectual property of the respective instructor or the university and are therefore protected by copyright law. This means that you may not upload or process such content in AI tools without explicit permission—even if it is only for study purposes, such as summarizing, transcribing, or translating.

This is especially important when using external AI services (e.g., ChatGPT, Gemini), as it is often unclear how your data is processed or stored. This creates a risk that protected content could be reused or unintentionally made public—something that may be problematic both legally and institutionally.

Even if the use of AI is intended to improve accessibility (e.g., transcribing a recording due to a disability), it is recommended that you discuss this in advance with your instructor or academic advisor.

Recommendations & Best Practices

• Ask your instructor whether you are allowed to input the material into an AI tool—many are open to supportive use if informed.
• Use only your own materials or anonymized notes if you want to be on the safe side.
• Use privacy-friendly AI tools when possible (see recommended AI tools).

Only if the use of AI tools is explicitly permitted for this assessment.

The decision whether you are allowed to use AI tools such as ChatGPT when preparing papers, essays, or other written assignments lies with the instructor or is defined in the relevant examination or module guidelines. AI tools are considered aids—therefore, their use must be clearly specified.

If no information about AI use is provided in the assignment description or module handbook, you should definitely ask in advance. This is important because using AI when it is not allowed may be considered academic misconduct, with corresponding consequences under §22 APSO or similar regulations.

If the use of AI is permitted, you are often required to document your use transparently, for example by:

• reflecting on it in the methodology section,
• including the prompts you used, or
• adding a note in the appendix of your paper.

Regardless of permission, the following applies: even if AI tools support you in writing, you must understand, reflect on, and take responsibility for the content yourself.

Recommendations & Best Practices

• When in doubt, always ask whether and how AI tools are allowed in your assessment.
• Clearly and transparently document your AI use if it is permitted—this fosters transparency and protects you.
• Do not use AI to bypass the work process, but rather as support for structuring, understanding, or reflecting.
• Avoid submitting fully AI-generated text without review—this contradicts the principles of academic work.

Whether you allow, restrict, or even require the use of AI should be guided by the intended learning outcomes of your course.

The decision of whether and how to use AI in teaching is complex and depends on various factors (including contextual conditions, potential risks to learning outcomes, fairness/equal opportunity, the relevance of AI in professional practice, and the verifiability of AI use). Instructors at TUM have considerable flexibility in this regard (see TUM AI Strategy). At the same time, this places responsibility on you as an instructor to make pedagogically sound decisions and to communicate them transparently to students.

In general, a distinction can be made between allowing AI in learning processes during the semester and in assessments. The following recommendations focus on the use of AI in summative assessments, as this is where instructors tend to have the greatest uncertainty. Formats particularly affected are those conducted partly or entirely without supervision. These include written assignments—from reports to academic papers—as well as portfolios and coursework. However, even for supervised in-class exams or midterm assessments, it is useful to consider the following questions and evaluate your assessment for “AI robustness.” Be sure to check your module description or the relevant program and examination regulations in advance to determine which assessment format applies in your case.

We recommend a structured three-step approach for making your decision, as outlined in the detailed document “How do I decide where to allow the use of AI tools in my teaching?”:

1. Clarify the intended learning outcomes of your course
2. Analyze your assessment tasks for AI robustness
3. Assess your students’ prior knowledge

General recommendations for implementation

• Communicate openly with your students about why AI tools are fully allowed, restricted, or completely prohibited. Explain which learning outcomes are intended in the course, how these rules help protect them, and why this matters for students.
• Consider establishing an AI commitment or learning agreement in which instructors and students jointly define how AI will be used responsibly.
• The goal is to foster a culture of trust, supported by well-designed assessments that ensure the validity of evaluation outcomes.
• Reliable monitoring of AI use is difficult—often impossible. If AI use is restricted or prohibited, assessments should be designed so that unauthorized use does not provide a decisive advantage.
• In many cases, it is important to explicitly train students in the competent use of AI, especially if this is a stated learning outcome of the course. Practice with your students should include, for example:
  • explaining how AI works, including its limitations and common errors,
  • identifying biases and misinformation,
  • critically evaluating and revising AI-generated outputs.

Communicate the permitted use of AI tools in writing and, ideally, multiple times (including verbally)—for example, in your Moodle course and during the first session of your class.

As an instructor, you can decide whether and how students are allowed to use AI in your course (see TUM AI Strategy; see also the question “How do I decide where to allow the use of AI tools in my teaching?”). It is important that you communicate your decision transparently and early on—ideally verbally in the first session and in writing on Moodle.

Formulate clear rules specifying for which activities AI tools are allowed and for which they are not (see sample wording below). Raise students’ awareness of potential pitfalls and the importance of transparently disclosing their use of AI (this should, of course, not be an advantage or disadvantage in grading). You may also want to demonstrate what appropriate use looks like—showing students how to use AI tools without negatively affecting their own learning process.

Sample wording:

• “The use of AI writing tools is permitted for brainstorming and idea generation. The final text must be written independently. The use of AI for automated text generation is not allowed.”
• “AI tools may be used in preparing the assignment if the prompt history and a reflection section on the use of AI tools are submitted along with the work. In addition, the AI tool used must be listed in the methodology section.”

Antwort wird generiert

Geben Sie Ihr Feedback zu BizChat ab.

Even though some of our answers touch on (data protection) legal aspects, they are intended as pedagogically oriented recommendations for the reflective and responsible use of AI in higher education. They do not constitute legal advice.

Please note that the use of AI tools in teaching, studying, and assessment is your own responsibility and must always comply with applicable examination regulations, data protection requirements, and copyright laws.

To help you better understand and contextualize the various responses, we have included the following two questions on AI and data protection to provide a brief, practical orientation.

If you are interested in a more comprehensive overview of legal aspects related to AI, we recommend consulting the AI Guidelines for Teaching in Higher Education.

Do not enter any personal, sensitive, or confidential data into AI tools – not even into systems approved by TUM.

Many AI systems, especially large language models (LLMs), process inputs on servers located outside of Europe. It is often unclear whether this complies with the General Data Protection Regulation (GDPR). Therefore, as a rule, no personal data (such as names, student ID numbers, email addresses, patient data, research participant information, etc.) or sensitive content (e.g., research results, exam data, internal information) should be entered into such tools. This applies even if the use of the AI tool has been approved by TUM.

Recommendations:

• Anonymize content before entering it (e.g., avoid real names, numbers, or project references).
• Use generalized or neutral examples or sample texts.
• Consult instructors or TUM’s data protection officer if you are unsure.
• Prefer data protection-compliant tools (see recommended tools).

Whenever possible, use AI tools that are hosted on European servers, have clear data protection policies, and—ideally—are approved by TUM.

Not all AI tools are equally suitable from a data protection perspective. Many providers, especially those based outside Europe, handle user data in unclear ways or store it for training purposes. As a student or instructor at TUM, you should therefore prioritize institutionally approved or privacy-friendly tools.

Recommended options:

• Microsoft Copilot (TUM campus license):
  TUM has entered into a campus licensing agreement with Microsoft. Copilot, integrated into Microsoft 365, provides a data protection-compliant framework through these contractual agreements (as of June 2025). More information on using Copilot at TUM can be found via the corresponding wiki link.
• Chat-AI (Academic Cloud):
  Another privacy-friendly option is Chat-AI from the Academic Cloud, which is operated on German servers by public universities.