ľ�ϸ���Ӱ��

Since 2 February 2025, the AI Act has prohibited certain AI systems and tools. Prohibited AI systems pose an unacceptable risk and may, for example, restrict people's freedom of choice, discriminate or manipulate, or can be in conflict with fundamental human rights. The AI Regulation does not replace the GDPR. Both laws apply.

The use of prohibited AI systems and tools in our university is therefore not permitted.

Which AI systems and tools are prohibited (Art. 5)?

• “social scoring” based on certain social behaviour or personal characteristics;
• “predictive policing” to assess or predict criminal offences committed by individuals;
• creating or (via scraping) expanding databases for facial recognition;
• manipulating or misleading people;
• remote biometric identification for law enforcement (with some exceptions);
• biometric categorization, whereby people are classified into certain sensitive categories on the basis of biometric data.
• Emotion recognition in the workplace and education (Art. 5.1.f.). Recital 18 of the AI Act explains that the term emotion recognition refers to emotions such as happiness, sadness, anger, fear, surprise, disgust, embarrassment, excitement, shame, contempt, satisfaction and amusement. It does not include physical states such as pain or fatigue. The mere detection of clearly visible expressions, gestures or movements is not prohibited, unless they are used to identify or infer emotions.

The Data Protection Authority (AP) supervises compliance with the AI Regulation and can impose fines in the event of violations.

The regulation of so-called high-risk AI systems and tools within the EU AI Act will take effect on 2-8-2026.

Most provisions in the AI Act focus on AI systems that pose a high risk to the health, safety or fundamental rights of individuals. These AI systems and tools must comply with a set of specific requirements laid down in the regulation to ensure their reliability and security.

1. The EU's AI Regulation classifies AI systems as high-risk AI if they meet one or both of the two conditions: The AI system is intended to be used as a safety component of a high-risk product (Annex I of the AI Act) or if the high-risk product itself is an AI-system. For example, when AI is applied in cars, medical devices and elevators.
2. The AI system falls under Annex III of the AI Act, which includes the education. Examples of a category of high-risk AI systems and tools are:
   1. AI systems or tools used for recruitment or selection of natural persons.
   2. AI systems or tools that are used to make decisions that affect the conditions around work.

More information can also be found on the website of the (AP).

Whether there is actually high-risk AI requires a careful assessment. It may be that the AI system or tool or use case falls e.g. under an exception.

Not covered by the AI Act:

• AI systems and tools used exclusively for military or defense purposes
• AI systems and tools used exclusively for scientific research and scientific development
• AI systems and tools purely for non-professional purposes use
• Open source AI systems and tools that do not function as a standalone AI system or are not used for regulated applications

Information about Low and minimal risk AI will be added to this website in mid-2025