Guidelines of the Johan Skytte Institute of Political Studies for specifying the use of generative artificial intelligence in teaching and learning

The University of Tartu does not lay down a single university-wide approach to the use of generative artificial intelligence (hereinafter: AI). The University has issued general guidance, but under the Academic Regulations the right to decide has been assigned to the level at which the relevant subject expertise lies, that is, to the teacher responsible for the course. Misuse of AI is dealt with under the procedure for handling cases of academic fraud in the Faculty of Social Sciences.

In formulating these principles, the Johan Skytte Institute of Political Studies has been guided, among other things, by strategic, ethical and didactic considerations. Our aim is that AI tools are used responsibly and ethically in the study process and that, in doing so, students achieve the intended learning outcomes. With AI, it is necessary to be aware of the risks it entails and to maintain critical thinking at every stage of its use. It is always the human being who is responsible for the end result of using AI.

At the Skytte Institute, each teacher determines, within their own course, the extent to which the use of AI is permitted in learning and assessment tasks. As a minimum, the teacher responsible indicates, in the course syllabus in the Study Information System (SIS), under Assessment methods and criteria, one of three categories for each task (use fully permitted / partly permitted / prohibited). The instructions for completing the task (which may be published elsewhere, e.g. in Moodle) must justify this choice at greater length, so that students can understand why this particular principle applies to that task.

• Use prohibited: AI may not be used when completing learning or assessment tasks.
• Use partly permitted: AI may be used to a certain extent when completing learning or assessment tasks.

• Use fully permitted: AI may be used extensively when completing learning or assessment tasks.

In tasks marked as “use prohibited”, students may not use AI tools at any stage of completing the learning or assessment task. This includes generating text, structuring it, producing summaries, translating, or using AI to rephrase what they have written themselves. The restriction does not cover spelling and grammar checking (e.g. the Word or Grammarly language checker). Where the task is one in which the use of AI is in any case not possible (for example, an exam taken in the classroom), this must also be indicated as such next to the task: “use prohibited (not possible)”.

This prohibition does not express any general opposition to technology or to AI on the part of the teacher; rather, it is a deliberate pedagogical choice: the learning outcomes of the course are best achieved when students rely entirely on their own thinking, reading and writing. In this context, the prohibition on using AI is intended to develop essential academic skills.

In tasks marked as “use partly permitted”, some AI applications and functions are permitted while others are not. The teacher decides where the line is drawn and, where necessary, sets the rules at the level of each individual task. This allows teachers to adapt the use of AI to the learning outcomes to be achieved on the course.

Permitting partial use is a deliberate decision, based on the principle that intellectually demanding tasks (e.g. constructing an argument, drawing evidence-based conclusions, synthesising sources) must remain the student’s own work. AI may support these processes, but it may not carry them out on the learner’s behalf.

In tasks marked as “use fully permitted”, students may use AI tools freely at all stages of completing learning or assessment tasks, including in drafting text, structuring it and generating content.

This freedom does not mean an absence of responsibility: the student remains the author of the work submitted and is therefore fully responsible for its accuracy, originality and integrity. Academic integrity, proper referencing and a critical attitude towards AI output are not merely required; they are crucial for the development of the learner’s professional competences, including AI competence.