of the EU the parliament and the council reached a preliminary agreement on the legislation on artificial intelligence.
The purpose of the legislation is to ensure, among other things, the realization of fundamental rights, democracy, ecological sustainability and the rule of law.
Parliamentary bulletin in the proposed artificial intelligence regulation, social scoring based on behavior and character traits and the use of manipulative artificial intelligence that aims to control people’s behavior and free will would be prohibited, among other things.
Biometric categorization of people’s sensitive data (for example, political, religious and sexual orientation) and the exploitation of certain groups of people’s vulnerabilities (for example, due to age, financial situation or disability) would also be prohibited in the regulation.
According to the regulation, the use of biometric identification systems would be possible for law enforcement authorities only in certain limited cases, such as finding victims of kidnappings or human trafficking, preventing an identified and immediate terrorist threat, and searching for suspects of certain crimes.
The entry into force of the Artificial Intelligence Regulation requires the approval of both the Parliament and the Council.
Planned according to the artificial intelligence regulation, creators of general-purpose artificial intelligence systems should comply with certain transparency requirements, which include, among other things, disclosure of the material used to teach artificial intelligence and compliance with EU copyright legislation.
The requirements are stricter for some general-purpose AI models with a wider impact.
Clear obligations would be imposed on artificial intelligence systems classified as high risk. High-risk systems would be defined, among other things, as artificial intelligence systems that try to influence voting behavior and election results.
Citizens would have the right to receive justifications for decisions made with the help of high-risk artificial intelligence systems that affect their rights.