Generative AI refers to systems that create new content based on user inputs (so-called prompts) — for example, text, images, audio, or video. The automated origin of this content is not always obvious at first glance.

→ Learn more: Self-paced course on AI in your studies, chapter “What is AI?”

A prompt is the input or instruction you give to an AI system to generate a response or content. Output quality depends greatly on how precise and well-thought-out your prompt is. Good prompting is a skill you can learn and practice.

→ Learn more: Self-paced course on AI in your studies, chapter “Working with chatbots”

Generative AI is based on statistical probability distributions rather than true language understanding. This means that systems produce sequences of words that are statistically likely, but not necessarily factually correct. Outputs can sound plausible and still contain false information. AI systems are therefore sometimes called “stochastic parrots” — they “parrot” likely word sequences without being able to verify their correctness.

→ Learn more: Self-paced course on AI in your studies, chapter “What is AI?”

Current generative AI systems have several fundamental weaknesses:

• They produce “hallucinations”, i.e., fabricated facts, quotes, or sources that sound convincing but are incorrect.
• They lack genuine contextual understanding and often do not reflect current developments, as they rely on training data from specific time periods.
• They exhibit bias: the selection and composition of training data can lead to content distortions with potentially discriminatory effects.
• AI systems generally do not transparently disclose their sources, which creates a risk that generated text may inadvertently contain plagiarism.

→ Learn more: Self-paced course on AI in your studies, chapters “What is AI?”, “AI and bias”, “Working with chatbots”

To evaluate AI-generated content on a specialist topic, solid subject knowledge is essential. Always question results critically and cross-check them with scholarly literature or other reliable sources. AI output is suitable as a starting point for further research, not as a replacement for it.

→ Learn more: Self-paced course on AI in your studies, chapter Documentation and disclosure of AI use

Yes. Academic AI is the AI platform of the Medical University of Innsbruck for students and staff. The platform is designed with privacy in mind, i.e., data remains on the platform unless explicitly stated otherwise. Academic AI includes various modules such as chatbot, document analysis, translation, and tailored AI (your own knowledge base), and offers access to language models from several providers (ChatGPT, Claude, Gemini, Mistral). Access is exclusively via your q-ID.

→ All information on Academic AI: www.i-med.ac.at/de/it/dll/academic-ai-fuer-studierende/

The Medical University of Innsbruck provides students with free access to Academic AI including a monthly token credit worth €5. Unless revoked, the costs are covered by the IT department.

The costs of using Academic AI are based on the number of tokens consumed. Tokens are the units AI systems use to process text — roughly corresponding to a word or part of a word. Every request uses tokens for both your input and the generated output. The more extensive and complex your prompt, the more tokens are used and the higher the cost. The choice of language model also affects consumption, as more powerful models generally have higher per-token costs.

Yes, you can find a comprehensive guide on Moodle.

AI queries are many times more resource-intensive than conventional search queries; both prompt complexity and model choice affect energy use. MUI’s Guidelines for Academic AI therefore call for responsible use: Before each query, consider whether it is truly necessary or whether a conventional search engine would suffice. Conscious AI use is not only good scientific practice but also a contribution to conserving resources.

→ Learn more: Self-paced course on AI in your studies, chapter “AI and the environment”

When you use AI in your studies, you must follow the Guideline on the use of AI systems in teaching and assessment at the Medical University of Innsbruck.

• Clarify in advance what is permitted in each course, and document your use transparently.
• All AI-generated content must be critically reviewed — this presupposes that you know the limitations and weaknesses of AI systems.
• Observe data protection and copyright, especially with regard to personal data, research data, and copyrighted material.
• Adhere to the principles of good scientific practice and always maintain your independence as an author.

→ Learn more: Self-paced course on AI in your studies, chapter Documentation and disclosure of AI use

AI can be a helpful study companion when used mindfully. It can support motivation, orientation, and new approaches. At the same time, there is a risk of outsourcing your own thinking or not building key competencies if you rely too heavily on AI. Whether AI supports or hinders your learning depends on whether you check suggestions, adapt them, and connect them to your own thinking.

→ Learn more: Self-paced course on AI in your studies, chapter “AI as a study companion”

Which generative AI tools you may use and to what extent is decided by the instructors in each course. AI use may be allowed in one course and explicitly prohibited in another. Always clarify in advance what is permitted and how to document any use. You remain responsible for everything you submit.

If AI use is permitted, it can support you with, for example, outlining, editing your language, or reflecting on questions. Ask yourself the following beforehand:

• What exactly do I want to use AI for?
• Is it allowed in this course?
• How will I document it correctly?
• Which tool fits my task?
• And how would I do this step without AI?

These questions help you decide deliberately and protect your independence.

You should not use AI if:

• it is prohibited in the course or your usage conflicts with course rules,
• you use it without transparent documentation,
• you fully rely on AI, i.e., hand over responsibility to a system.

Do not use AI in ways that hinder your learning process or jeopardize your independence as an author. Do not use AI as a shortcut, but as a tool — with a clear purpose.

No. Do not enter personal data, patient data, confidential research data, or copyrighted material into AI tools. This also applies in Academic AI: Think carefully about what data you share before every input. Sensitive information does not belong in AI systems.

No. It may be tempting to see AI as a tool that relieves you of tedious writing. This is only partly true, as revising AI-generated text can take a lot of time. You may also miss important learning opportunities and only partially develop competencies by relying on AI.

Unauthorized use of ChatGPT or similar AI systems in the context of (e-)exams (e.g., KMPs/iKMPs/MCQs) and other graded coursework in a class constitutes obtaining performance by fraudulent means. This results in the nullification of your performance. Examinations whose grading is declared null and void count towards the total number of permitted resits.

Always observe the instructions of your instructors and the legal provisions (see in particular the “Guideline on the use of AI systems in teaching and assessment at the Medical University of Innsbruck”, Sections 73 para. 1 no. 2 and 89 para. 1 lit. c Universities Act, Section 2a para. 3 no. 2 Higher Education Quality Assurance Act).

The decision as to whether AI is permitted is made by the respective instructor. If AI systems are used as part of written (partial) assessments or when producing academic or other written work, the guideline requires the following statement in the declaration of authorship:

→ Learn more: Guideline on the use of AI systems in teaching and assessment & self-paced course on AI in your studies, chapter Documentation and disclosure of AI use

AI-generated text is not automatically plagiarism in the classical sense, as it is not copied directly from an identifiable source. Nevertheless, AI-generated text can contain passages that closely resemble existing works without being marked as such. Furthermore, if you present AI-generated text as your own without disclosure, you violate the principles of good scientific practice and MUI’s guideline.

→ Learn more: Self-paced course on AI in your studies, chapter “Law and ethics”, Documentation and disclosure of AI use