AI in Teaching – ETF Guidelines

General Principles

  • The ETF recognizes the growing importance of AI. It encourages professors and students to acquire the skills necessary for a meaningful and responsible use of AI tools.
  • The ETF asks its professors to integrate appropriate AI tools into their teaching in order to familiarize students with their possibilities, limitations, and risks.
To top

Risks and Recommendations for a Critical Use

In addition to their advantages, AI tools also pose risks, which can be avoided through careful use.

  • Limitations of AI. Common AI applications are chatbots, which are very good at formulating, structuring, and arguing. They can recognize patterns and apply them to (textual) examples. Their responses are based on statistics. Chatbots do not possess actual knowledge, they cannot think, and they lack personal experience. For complex theological debates, new scholarly theses, and texts with individual style and depth, so far, they are insufficient. AI tools are only as effective as the people using them.
  • Misinformation. AI tools tend to hallucinate. Correct information is often mixed with incorrect information, which is only noticeable to people well informed. Therefore, information generated by AI tools should not be trusted blindly. Instead, it must always be critically verified.
  • Manipulation. Like information in general, AI-generated content includes value judgments (biases). To avoid the risk of being manipulated, it is essential to critically reflect on judgments in AI-generated information.
  • Data protection violations. Data entered into AI tools is generally not protected and may be reused. Users must not input personal data into AI tools, and caution is advised with sensitive information. The University of Vienna provides lecturers and students with free access to Microsoft 365, which includes Copilot (GPT-4). Here, data is protected through “Enterprise Data Protection.”
  • Decline of one’s own skills. AI tools are capable of solving even complex tasks and producing well-formulated (but often superficial) texts. This can be practical because it saves time. However, it includes the risk that users’ thinking and writing skills are not sufficiently developed and will deteriorate. Independent thinking (about texts, problems, questions, etc.), research, and writing are essential components of learning and intellectual growth. Delegating these tasks to AI tools too often results in missed learning opportunities and can lead to dependence on AI, resulting in a loss of the ability to critically evaluate it.
  • Energy consumption. The development and use of AI require enormous amounts of energy. For sustainability reasons, its use should be kept to a minimum.
To top

Rules for the Practical Use of AI at the ETF

  • AI tools as aids. In assessments, the use of AI tools is permitted only as an aid, if at all, (e.g., for brainstorming), but not as a substitute for one’s own work.
  • Write your own texts. Copying entire passages generated by AI tools is considered cheating; it constitutes a form of plagiarism.
  • Text correction. Unless explicitly prohibited by professors or supervisors, correcting one’s own texts using programs such as DeepL is permitted. Students are encouraged to carefully review suggestions for corrections so that they can improve their own writing abilities without AI in the future.
  • Documentation of use. If AI tools are used as aids in assessments, this use must be documented—comparable to the use of secondary literature. The AI tools used must be listed in the bibliography under aids, and footnotes must include information which ideas were generated with the help of AI.
  • Not a substitute for citations. Referring to AI tools does not replace standard academic citations. Important information must be supported by scholarly publications.
  • Exams. Professors are asked to design assessments in a way that minimizes the risk of cheating with AI tools.
  • Cheating. If professors or supervisors suspect that students have used AI tools illegitimately, they may invite them to a clarifying talk. If the suspicion is confirmed, this will be noted in the academic records.

Professors define the details of legitimate use. Professors and supervisors determine the extent to which AI use is permitted in their courses and supervised work. Students are asked to consult them directly in case of uncertainty.

To top

Further Information

The University of Vienna has compiled extensive information on the use of AI in teaching; the ETF recommends that students and lecturers consult these resources:

Adopted at the faculty meeting on January 20, 2025

To top