University of Ljubljana Sets Common Rules for AI Across Campus
By Suad Seferi ·
The University of Ljubljana has introduced university-wide guidelines for artificial intelligence, creating a common framework covering students, lecturers, researchers and professional staff. Published at the beginning of September, the guidelines apply across learning, teaching, assessment, research, development work and university administration. Rather than banning generative AI or leaving individual faculties to deal with it independently, Slovenia’s largest university is setting common principles for how the technology should be used. Those principles include human responsibility for the final result, protection of personal and confidential data, transparency when AI is used, academic integrity, equitable access and using AI only where its role is proportionate to the task. The university has also created separate guidance for different groups, including students, teaching staff, researchers and professional employees. The human remains responsible One principle sits at the centre of the framework: responsibility cannot be transferred to an AI system. The university says the person using AI remains responsible for the final result and that AI cannot replace professional, pedagogical or research judgement. That has practical implications across campus. A student cannot simply point to an AI tool when information in an assignment is false. A researcher remains responsible for the accuracy and integrity of work produced with AI assistance. A lecturer using an AI system as part of assessment still carries responsibility for the academic decision. The same principle applies to administrative work. The university’s guidance therefore treats AI as a tool that can contribute to a process rather than as an independent decision-maker. One framework for an entire university Universities across Europe have responded to generative AI in very different ways. Some initially introduced restrictions around ChatGPT and similar systems. Others allowed individual faculties or lecturers to create their own policies. That can create uncertainty for students moving between courses where acceptable AI use differs considerably. Ljubljana is taking a more centralised approach by establishing minimum rules across the whole university while allowing them to be applied differently depending on context. The guidelines specifically cover teaching and assessment but extend further into research, professional activities, preparation of university materials and digital services containing AI functionality. They do not replace existing university rules covering academic integrity, research ethics, personal data, copyright or information security. Instead, the AI framework sits alongside them. That is an important distinction. Many of the risks created by AI are not entirely new. Entering confidential research data into an external AI platform, for example, may already conflict with data-protection or research rules regardless of whether an AI-specific policy exists. AI literacy rather than prohibition The university’s approach also reflects a shift in how higher education is responding to generative AI. The question is increasingly moving away from whether universities can prevent students from using AI. The more difficult issue is teaching students and staff when its use is appropriate, how to verify what it produces and when relying on it undermines the purpose of an academic task. University of Ljubljana material accompanying the new guidelines makes that position explicit. Human judgement remains central, while AI can provide assistance, warnings or alternative perspectives without taking over the final decision. That approach becomes particularly important in assessment. There are situations where allowing AI may be useful because students need to learn how to work with the technology. There are also tasks where independent reasoning is precisely what is being assessed and AI use would defeat the purpose. A single rule cannot cover both cases. A model other regional universities may watch For universities elsewhere in the Balkans, Ljubljana provides a practical example of how an institution can move beyond informal recommendations. The framework does not attempt to predict every AI tool that students or researchers might use. Instead, it establishes principles that can continue to apply as the technology changes. That is deliberate. The University of Ljubljana describes the guidelines as a living document that will be updated as technology, legislation and the needs of the university community develop. That may prove more durable than policies tied to individual products such as ChatGPT, Gemini or Claude. AI tools will change. The harder questions around responsibility, transparency, privacy and academic judgement are likely to remain. For Ljubljana, those questions now have a university-wide framework rather than being left entirely to individual classrooms.