Implications of using artificial intelligence in educational systems

5. Artificial Intelligence and the Legislative Framework in Higher Education: What does the European Artificial Intelligence Act (EU AI Act) introduce?

The European Act on Artificial Intelligence(the EU AI Act) is the first comprehensive legal act of the European Union to regulate AI systems within the common market framework. 
In terms of higher education, this means that institutions and faculties introducing AI tools and systems must assume a new level of responsibility: systems that can affect student outcomes, selection, or progression may be classified as “high-risk.” The EU AI Act introduces a categorisation of AI systems based on their potential risk: minimal, high-risk, and unacceptable.

For example, systems that automatically select candidates for admission or assign grades can fall into the high-risk category and must meet specific requirements.

Higher education institutions have several key steps ahead:

• Conduct a risk assessment and test AI tools before full implementation.
• Ensure that the data driving the system meets the criteria for quality, relevance, and initial representativeness.
• Ensure transparency: students and teachers must be clearly informed about the system’s role and limitations.
• Introduce human oversight of results that may affect students’ academic status.
• Ensure the protection of personal data in accordance with the General Data Protection Regulation (GDPR) and protect against unauthorised disclosure.

Rights of students and stakeholders in the educational process

Students and other users of the system have the right to:

• Clear and timely information about how AI is used in their education.
• Understand the basis on which the system makes assessments or recommendations.
• The possibility to object and request human review of decisions mediated by AI.

If a higher education institution uses an AI system to automatically grade student work, in order to comply with the EU AI Act, it must:

• Document the system’s reliability, including testing on a variety of data.
• Inform students about the grading process, algorithm, and appeal options.
• Ensure human oversight of final grades to prevent unfair or non-transparent decisions.

Implementation of the legislative framework (EU AI Act) and timeline

The EU AI Act entered into force on August 1, 2024, with the implementation of individual provisions in phases. Higher education institutions must prepare for the gradual implementation of this legislative framework and adapt their systems and practices accordingly.

The application of AI in higher education can be a powerful tool for improving learning and administration, but its success depends on the responsibility of all stakeholders. In accordance with the EU AI Act, higher education institutions must establish standards that guarantee safe, transparent, and fair use of AI, with human accountability.

Reflective questions:

• What artificial intelligence tools are you already using in your institution, and how do you plan to align them with the EU AI Act?
• How would you establish a procedure for human review of decisions made using artificial intelligence?
• What is the first step you would suggest to improve transparency in the use of artificial intelligence in the classroom?