Guidelines on the Use of Artificial Intelligence at the Faculty of Geo- and Atmospheric Sciences
Prepared by Emily Collier, Alexander Gohm, Jutta Kister, Gertraud Meißl, Elisabeth Militz, and Jeffrey Munroe
(Generative) artificial intelligence (AI) can support studies and research in many ways, for example with literature searches, brainstorming, generating codes, working with texts (summarizing, translating, refining), or creating figures and images. However, the purpose of university study is to develop skills and competencies, not to simply acquire knowledge. Independently engaging with problems, working through tasks on your own, and developing solutions are essential to the learning process and for developing competencies. When thinking and writing processes are outsourced to AI-based automation, the ability to independently, creatively, and innovatively accomplish tasks is lost, while human language, perspectives, and ways of thinking may become increasingly homogenized. You should therefore critically assess where AI tools could provide useful support, and avoid situations where AI significantly hinders the learning process.
1. Good Scientific Practice and the Use of AI
Even in the age of artificial intelligence (AI), the fundamental principles of good scientific practice (https://www.uibk.ac.at/de/forschung/qualitaetssicherung/gwp/) apply without restriction. The use of (generative) AI never exempts authors from carrying out their own original scientific work. The generation of entire passages or chapters without an author’s own draft is not permitted and is considered academic misconduct. Ultimate responsibility for all content, scientific interpretations, and research data lies entirely with the authors and cannot be delegated to an AI tool.
In accordance with principles of honesty, transparency, and reproducibility, the use of AI tools must be fully documented (see Section 5). Failure to disclose the use of AI is considered plagiarism. Furthermore, because AI-generated outputs are not verifiable primary sources, are often not exactly reproducible, and carry the risk of hallucinations (“fake sources”), AI systems must never be cited as sources of knowledge. AI tools also cannot be listed as co-authors.
Intellectual property rights and fidelity to original sources must be respected. The use of AI for literature searches or synthesis does not exempt authors from the obligation to independently obtain, critically analyze, and directly cite the underlying primary literature. AI-generated statements must always be checked against the original sources in order to ensure that scientific arguments rest on a sound and verifiable foundation.
2. Data Protection and Copyright
Data protection and copyright must be strictly observed. The relevant legal foundations are provided by the EU AI Act, the European General Data Protection Regulation (GDPR), the Austrian Data Protection Act (DSVO), and copyright law. In particular, it is prohibited to upload recordings and photographs from lectures, materials provided by instructors or fellow students, unpublished research data, etc. into AI models (for example, in order to generate summaries, flashcards, or examination questions). Such practices constitute a serious violation of copyright law. In addition, for data protection reasons, personal data must not be uploaded to AI systems.
3. Ethical Aspects
AI tools are embedded in, arise from, and exacerbate planetary and social inequalities. Economic value creation through artificial intelligence is likewise distributed unequally around the world. Furthermore, the development and use of AI tools are associated with enormous environmental and social costs. Some AI data centers are constructed on territories from which Indigenous populations have been violently dispossessed, and they consume large and ever increasing amounts of electricity, which is predominantly generated from fossil-fuel and nuclear energy sources. Increased energy consumption leads to rising electricity prices, which disproportionately affect poor and marginalized populations. High demand for electricity is also associated with high water consumption for cooling.
It should therefore be recognized that even meaningful uses of AI tools in education and research carry costs and consequently reinforce and perpetuate global inequalities and the climate crisis.
4. Use of AI in Courses
Course instructors decide whether AI tools may be used as permitted aids in their courses, and how they may be used in different contexts, for example, in writing, programming, examinations, or other coursework. When developing these rules, instructors should consider the learning outcomes and assessment methods of their courses, as well as issues such as data protection and copyright. Instructors must clearly communicate their specific policies regarding the use of AI at the beginning of each course, including any requirements for documentation. In addition, instructors should discuss the advantages, disadvantages, and ethical implications of using AI within the context of their particular course, especially in seminars, practical classes, and field excursions. Given the rapid evolution of AI, instructors are encouraged to stay informed on this topic and to regularly review and, where necessary, revise their course policies. For written theses, the guidelines described in the following section apply accordingly.
5. Use of AI for Theses
Theses are fundamental demonstrations of students’ ability to conduct independent scientific work. The requirements described in the preceding sections must therefore be fully observed. A declaration must also be included when the thesis is submitted.
The thesis supervisor may restrict the use of particular AI tools or particular forms of AI use, for example for editing text, diagrams, or code.
If AI tools are used, their use must be clearly, transparently, and comprehensibly disclosed. An AI declaration must be included at the beginning or end of the thesis, even if no AI tools were used. This declaration must transparently and thoroughly identify the AI tool used, the purpose for which it was used, and the measures taken for quality assurance and risk control (examples are provided in the Appendix).
If AI tools are used during the research process, for example, in the processing, analysis, or interpretation of research data, this must also be explained transparently in the Methods section of the thesis, in a separate subsection with its own heading if appropriate. It must always be clearly distinguishable which parts of the work are based on the author’s own efforts and which parts were supported by AI.
The author is always responsible for preparing the thesis with the necessary care. Maintaining a log is recommended to facilitate documenting how AI tools were utilized throughout lengthy research and writing processes. To provide evidence in case of future inquiries, it is particularly recommended that a record of prompts be retained (see also the University of Innsbruck’s “Guideline for Safeguarding Integrity in Academic and Artistic Study, Teaching, and Research,” published on 13 July 2026 in the university announcements for the 2025/26 academic year, Issue 133).
Appendix: Possible Versions of the AI Declaration
Version A (Text)
Declaration on the Use of AI Tools in Theses
Instructions: If you used generative AI tools (e.g., ChatGPT, Claude) while preparing your thesis, please enter the relevant information in the brackets below and delete the bracketed text that does not apply. This declaration must be formatted as a separate section and placed either at the beginning of the thesis, or at the end before the reference list.
Template Text
I hereby declare that generative artificial intelligence (AI) tools were used as aids in the preparation of this thesis. Specifically, [insert name of tool] was used to provide support with [e.g., data analysis / programming in Python/Fortran / formatting LaTeX code]. In addition, [insert name of tool] was used for [language support, including checking grammar and wording, improving the readability of text written by the author].
I hereby declare that I take full responsibility for the content of this thesis. The scientific arguments and interpretation of the results were developed independently by the author. All AI suggestions were critically reviewed, revised, and adapted by the author before being incorporated into the thesis.
Example of a Completed Declaration
The following example illustrates how the template may appear when completed by a student:
I hereby declare that generative artificial intelligence (AI) tools were used as aids in the preparation of this thesis. Specifically, ChatGPT and Claude were used to support exploratory data analysis, review WRF Fortran source code, and develop Python scripts for graphical visualization. In addition, these tools were used for language support, including checking grammar and wording, and improving the clarity of text originally written by me.
I hereby declare that I take full responsibility for the content of this thesis. The scientific arguments and interpretation of the results were developed independently by the author. All AI suggestions were critically reviewed, revised, and adapted by the author before being incorporated into the thesis.
Version B (Table)
Please also always consult the University of Innsbruck FAQs on the topic of AI: https://www.uibk.ac.at/de/universitaet/digitalisierung/ki-uni-innsbruck/faqs/
Further Reading:
Aczel, M., Chamanara, S., Matin, M., Farsi, A., Marwala, T., Madani, K. (2026). Environmental Cost of AI's Energy Use: Carbon, Water and Land Footprints. United Nations University Institute for Water, Environment and Health (UNU-INWEH), Richmond Hill, Ontario, Canada. doi: 10.53328/INR26RMA002
Sourati Z, S. Ziabari A, Dehghani M. (2026): “The homogenizing effect of large language models on human expression and thought.” Trends in Cognitive Sciences, 2026, 30, 805–816. doi: https://doi.org/10.1016/j.tics.2026.01.003