UL approves regulations for use of AI tools – clear boundaries in digital age

With this decision, the University of Latvia becomes one of the first higher education institutions in Latvia to regulate the use of AI in the study process at the Senate level, in accordance with the amendments made to the Law on Higher Education Institutions in January. The regulation clearly defines both the opportunities and the boundaries at a time when generative AI tools are rapidly entering education, the labour market, and everyday life of society.

“As a research university, we believe it is our responsibility to follow technological developments and make use of the opportunities they provide, while critically assessing the associated risks and implementing the necessary preventive measures. The University does not prohibit technologies; rather, it promotes their responsible use. Our goal is to ensure a study process whereby students genuinely acquire the promised knowledge, skills, and competencies. Students must be informed about both the possibilities and the limitations of AI tools – in studies, technology should serve as an assistant, not a substitute,” says Kristīne Strada-Rozenberga, UL Vice-Rector for Studies.

This step enhances the position of UL as a responsible and innovative higher education institution that both promotes the integration of technology into the study process and safeguards the principles of academic integrity, data security, and ethics.

The regulations stipulate that the use of AI at UL is based on four core principles: integrity, transparency, responsibility, and data security. Students will be required to clearly indicate the use of AI in their work, while teaching staff will be responsible for defining in which cases and to what extent the use of AI is permitted within a specific study course.

Unauthorized use of AI tools will be treated as the use of unauthorized aids and will be considered a violation of academic integrity.

The regulations place particular emphasis on the protection of personal data and intellectual property. Students and teaching staff will be prohibited from uploading restricted-access information or study materials to AI tools for which UL has not purchased a license or that have not been officially implemented at the University. The creation of deepfakes and the use of unlicensed virtual AI assistants on behalf of the University are also prohibited.

At the same time, the regulations ensure an important balance – the results of generative AI content detection tools may not be used as the sole basis for evaluating a student’s work or as evidence of a violation of academic integrity. Final responsibility for the assessment will always rest with the lecturer.

The regulations provide for a flexible approach – the permissibility of AI use within a specific course will be determined by its teacher, taking into account the study objectives and intended learning outcomes. Faculties and study programme directors will be able to further specify the procedures for AI use in accordance with the specifics of their field.

To support the implementation of the regulations approved by the senate, the updated Guidelines for the Use of Artificial Intelligence at the University of Latvia – currently in their second version – will be introduced. They compile recommendations and examples of good practice for the use of AI in the study process.

The recommendations described in the guidelines are based on an analysis of the approaches taken by leading universities worldwide, including Oxford, Cambridge, Harvard, and the Massachusetts Institute of Technology, as well as on reports by international organizations and institutions (OECD, UNESCO, EUA, etc.), regulatory frameworks (including the European Artificial Intelligence Act adopted by EU in 2024), and relevant research literature.

The implementation of an AI environment at the University of Latvia is one of the activities supported by the Recovery and Resilience Mechanism project 'Internal and External Consolidation of the University of Latvia' (No. 5.2.1.1.i.0/2/24/I/CFLA/007).