This page provides students, doctoral candidates, faculty, teachers and staff with key information on teaching and learning with artificial intelligence (AI) at Bauhaus-Universität Weimar.
University-licensed AI Tools
Bauhaus-Universität Weimar offers its members access to software licences and AI tools that are available via API interfaces on the university’s own IT infrastructure or on the servers of the IT Centre for Thuringian Universities.
The following AI applications are accessible with university login via Shibboleth or freely available:
| AI Tool | Purpose |
|---|---|
| Chat AI | Generate text using various models such as Ilama, Deep Seek, Mistral and Qwen |
| Open Model Lab | Generate text using Gemma |
| HS-ITZ Übersetzer | Translate texts with DeepL |
| ScopusAI | Academic source research |
| Statista Research AI | Retrieve statistical data |
| HS-ITZ AI-Chatbot | For staff only: Generating text with ChatGPT |
Further information on the SCC’s AI-based services can be found in the Service Catalogue under the heading »Artificial Intelligence«.
Frequently Asked Questions (FAQ)
What is generative AI?
»Generative AI« refers here to computer models that generate new content (including text, images, and music) based on human or machine input. The models rely on patterns and structures that they have learned from a training data set.
The term »AI applications« refers here to digital tools that allow the general public to interact with generative AI models through user-friendly interfaces, such as web interfaces (API, HAWKI) or service providers like ChatGPT, Claude.AI, or Leonardo.AI or the university’s own HAWKI interface.
Am I allowed to use AI?
Yes, generally speaking, the Bauhaus-Universität Weimar encourages its staff and members to engage critically with and use AI applications responsibly.
The »Guidelines on the Use of Generative Artificial Intelligence« are intended to provide guidance in this regard. They are advisory in nature and are not binding. They offer a clear framework for the responsible, creative and critical use of generative AI. Binding regulations (e.g. for examinations) are governed by the respective studies and examination regulations.
In the context of examination work, such as the preparation of seminar papers and final projects, however, AI applications should only be used as a supporting tool and with the express permission of the examiner (»Declaration of Release«). In principle, independent work should continue to predominate. The use of AI must be transparently documented in the context of examination work and academic research (list of resources and »Declaration of Originality«).
How do I cite/reference AI-generated content?
AI-generated content must be clearly documented in the context of academic research and examination papers.
We recommend listing the use of AI as a full reference in the list of resources/bibliography at the end of the paper:
- ChatGPT 4.0; OpenAI; openai.com/chat
Wording assistance and text correction, »Full prompt«, AI-generated, 10 March 2025
In addition, AI-generated sections and graphics should be accompanied by a brief citation at the relevant point in the text (in the text, as a footnote or in the caption):
- AI-generated using ChatGPT 4.0, OpenAI, 30 January 2025, chatgpt.com
The exact form of documentation is to be agreed upon in the »Declaration of Release« between students and academic staff or examiners. A full citation including prompts is required, for example, if the AI has generated its own analytical content or if the AI-supported process itself is the subject of the examination.
Can AI detection software be used to identify AI-generated content?
AI detection software is currently unable to reliably identify AI-generated content; therefore, its use is not legally sound and is consequently not recommended.
Which examination formats are best suited to preventing cheating through the use of AI?
To prevent attempts at cheating or to rule out the use of AI, the following examination formats may be used:
- Oral examinations and colloquia
- Face-to-face formats (written examinations, presentations on site)
- Examination formats focusing on individual reflection and application
- Open-book formats
- Supplementary discussions on seminar or project work
What risks should I be aware of when using AI?
Bias, Fairness, and Equal Opportunities
AI systems can reproduce prejudices that exist in the training data, leading to discriminatory results. Avoiding this requires not only university-wide awareness, but also careful review of generated content by individual users. To support and promote equal opportunities, the Bauhaus-Universität Weimar offers free access to various AI applications. Additional training is offered to improve AI competencies and raise awareness of inequalities.
Data Protection, Safety and Accountability
Generative AI processes massive amounts of data. This poses a risk to data protection and the security of sen- sitive research and administrative data. Universities must implement effective data protection measures, assign clear responsibilities to protect sensitive information, and ensure accountability and transparency.
Employment, Qualifications, and Right of Co-Determination
While generative AI can make certain work processes easier, it also raises questions regarding job security, espe- cially in administrative and support functions. At the same time, the demand for university members and affili- ates to skilfully use generative AI in a sensitive and responsible way is increasing. This requires targeted training and adapting teaching methods in order to prepare students for a working world that is increasingly dominated by AI. Employees also need to be involved early on in decision-making processes in order to ensure acceptance and develop fair solutions.
Transparency and Accountability
Generative AI generates output based on statistical patterns and probabilities; it does not make independent decisions and cannot consciously understand content. This can lead to uncertainty in art, research, teaching, and administration, especially if AI-generated content is not clearly identified as such. In order to establish trust in AI-supported processes, it is essential to disclose how they function and where they are being used. Clear rules must be agreed upon for their use. Section 4 includes recommendations on how to achieve this.
Sustainability and Resource Use
Training AI models and operating data centres use considerable amounts of energy and water. In line with its sus- tainability strategy, the Bauhaus-Universität Weimar is working on the responsible use of computing resources. Before each use, users should ask themselves whether their question can be answered without generative AI.
Plagiarism and Academic/Artistic Integrity
AI tools can generate texts, images, music, or programme codes where the source is uncertain. This is especially true if sources have not been correctly cited or if content has been automatically compiled from existing mate- rials. Students, instructors, researchers, and artists can unintentionally pass off the ideas of others as their own. The makes it difficult to assess independent work.
Hallucinations and Reliability of AI Information
AI applications can generate erroneous or made-up content that appears plausible at first glance, but is false or misleading in terms of its content. This poses major risks for scientific or academic work if unreliable or un- verified sources are used. Uncritical use of AI can also lead to false conclusions in administrative and academic decision-making processes. It is therefore essential that AI-generated content is always reviewed, critically scru- tinised, and transparently identified.
Reducing Information and Language Complexity
Generative text models tend to produce unnecessarily complex sentences. As a result, reading these sentences requires more attention from the reader, distracting from the essentials. Text models must be explicitly prompted to generate simple language.