OUR PUBLICATIONS

A New Era with the European Union Artificial Intelligence Regulation: Implementing the General Provisions and Prohibitions


11 February 2025 - Av. Melda Şener, LL.M.

A New Era with the European Union Artificial Intelligence Regulation: Implementing the General Provisions and Prohibitions

The European Union's ("EU") Artificial Intelligence Regulation ("AI Act") will be gradually implemented as of February 2nd, 2025. This is the beginning of a new era for the EU and the rest of the world. To recall briefly, the AI Act categorizes the use of artificial intelligence systems according to their risk level, prohibits some activities entirely, and imposes general obligations. As AI technologies are increasingly integrated into our lives, regulators must manage the risks in this area. With the AI Act, the EU aims to protect users and provide the framework to support innovation.

General Provisions

Although the prohibited activities and related discussions are prominent in the AI Act, as of February 2nd, 2025, the implementation of the general regulations, which are included as four articles in Part I, has officially started. Among these, we can summarize the highlights as follows:

  • Subject and Scope of the AI Act: The AI Act aims to ensure that artificial intelligence systems are developed, marketed, and used safely and transparently without harming human rights, health, and the environment, as well as to support technological developments. Various monitoring and oversight mechanisms can be used to achieve these objectives, such as banning or strictly regulating specific systems and regulating transparency rules. The AI Act covers anyone who carries out activities related to artificial intelligence systems in the EU, regardless of where they are headquartered.
  • Definition of Artificial Intelligence: According to the guidance published by the European Commission on February 6, 20251, the AI Act only applies to systems that meet specific criteria. According to the published guidance, an artificial intelligence system is defined as a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Understanding this definition is critical to correctly interpreting the scope of regulation. The guide provides essential information for the correct interpretation of important issues within the given definition.
  • Other Definitions: Part I, Article 3 contains definitions of concepts prominent under the AI Act and in practice. As the AI Act becomes more widely used, the correct interpretation of definitions will become even more important.
  • Informing and Artificial Intelligence Literacy: The parties involved in producing and introducing artificial intelligence (operator) must provide the necessary information to their employees and users for the correct and safe use of the artificial intelligence system. It is useful to
 
   
 

 

1 To access the guide: https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-ai-system- definition-facilitate-first-ai-acts-rules-application

 

understand the scope of this information correctly. It is possible to refer to a program designed by considering the technical knowledge, experience, and education level of the people to be trained, as well as the content of the relevant artificial intelligence system, rather than a general and abstract information transfer. Correctly implementing such a program can also feed transparency efforts and create benefits.

Prohibited Activities

The AI Act prohibits specific artificial intelligence applications starting February 2nd, 2025. These bans are introduced to protect the fundamental rights of individuals and prevent unethical artificial intelligence practices. Let's take a brief look at the bans and their highlights below:

  • Subliminal Manipulation: The use of artificial intelligence systems that alter the behavior of users without their conscious awareness is prohibited. Such systems are not permitted where they create a potential for harm by materially influencing consumers' subconscious minds and leading them to make decisions they do not want to make. Likewise, systems that materially influence disadvantaged or vulnerable social groups with such artificial intelligence systems and create the potential for harm are also prohibited.
  • Social Scoring: Systems that rate individuals by assessing their social behavior and economic status are prohibited from being used in such a way that the treatment of such individuals results in outcomes that are unfair or offensive
  • Criminal Risk Assessment Systems: It is prohibited to use artificial intelligence to predict the risk of offending based solely on personality traits or profiling. However, artificial intelligence systems based on human assessment, directly verifiable, and supporting objective crime-related data are excluded from this scope.
  • Expansion of Facial Recognition Databases: It is prohibited to use artificial intelligence systems that create or expand facial recognition databases by collecting random facial images from the Internet or security cameras.
  • Emotional Recognition Systems: Artificial intelligence systems that analyze individuals' emotions are banned in education and workplaces. This ban was introduced specifically to prevent the monitoring of employees' and students' psychological states.
  • Biometric Categorization Systems: The use of artificial intelligence systems that categorize individuals based on their biometric data into sensitive categories such as race, political opinion, trade union membership, religious or philosophical beliefs, and sexual orientation is prohibited. However, the filtering of lawfully obtained biometric data or categorizing biometric data by law enforcement agencies is excluded from this scope.
  • Real-time Remote Biometric Recognition: Real-time biometric surveillance systems in public areas are restricted, with specific legal exceptions. It is stated that police and security units may utilize these systems in exceptional circumstances within certain legal frameworks. It should be acknowledged that the extent to which such exceptions will be applied fairly and justly will be determined in practice.

Conclusion

 

The EU AI Act aims to ensure that AI technologies are developed in accordance with human rights, ethics, and the rule of law. The bans and general obligations that have started to be implemented are essential for all actors operating in the artificial intelligence ecosystem. Although the articles regarding sanctions and fines have not yet been implemented, operating in compliance with the AI Act as of this period is an investment in the future. Developing AI applications in an auditable, transparent, reliable, and ethical manner will both minimize legal risks and increase user confidence.