AI Act

The AI Act was published in the Official Journal of the EU on 12 July 2024. It regulates the development and use of artificial intelligence (AI) and sets high standards for safety, transparency and accountability to protect the fundamental rights of EU citizens. It follows a risk-based approach, whereby applications with higher risks are more strictly regulated or even prohibited. The AI Act will enter into force on 1 August 2024 and will apply from 2 August 2026. However, specific parts of the AI Act will become effective earlier, while others will not take effect until later. Therefore, it is important to become familiar with the regulations to determine whether one is using systems that are no longer permitted under the AI Act (prohibited AI practices).

^{This infographic was created with the help of AI.}

Loosely defined, an AI system is a computer system that imitates human abilities such as logical thinking, learning, planning and creativity. AI systems can also perceive and interact with their environment through cameras, microphones and sensors.

Whenever a "new" and "innovative" computer system is to be implemented which is capable of self-improvement and has human-like abilities, it must be examined whether it qualifies as an AI system under the AI Act.

AI is a cross-sectional matter that is not regulated solely be the AI Act but cuts across several fields of law, such as:

• labour law (shop agreements, personnel selection process, etc.);
• regulatory law (Banking Act, Insurance Supervision Act, etc.);
• consumer law ( Product Liability Act, transparency, fairness);
• data protection law;
• IP law; and
• fundamental and human rights (prohibition of discrimination, freedom and security).

Generally, supportive AI systems must be used with caution, as errors in their advice or forecasts could result in claims for damages.

The AI Act imposes different obligations on different actors. The key obligations include:

• information obligations;
• transparency and documentation obligations;
• protection of personal data;
• security and liability.

• Up to EUR 35m or (if the offender is an undertaking) up to 7 % of its total worldwide annual turnover for the preceding financial year, whichever is higher, for non-compliance with the prohibition of AI practices.
• Up to EUR 15m or (if the offender is an undertaking) up to 3 % of its total worldwide annual turnover for the preceding financial year, whichever is higher, for non-compliance with certain obligations under the AI Act, including non-compliance with the relevant provisions on general-purpose AI models.