How the European Union wants to tame artificial intelligence

The planned rules for artificial intelligence (AI) in the European Union are taking shape: MEPs in the European Parliament in Strasbourg voted yesterday, Thursday, for conditions for the use of AI systems such as the chatbot ChatGPT, which can formulate texts in a matter of seconds.

84 votes in favor, seven against, twelve abstentions: the decision of the parliamentary committees for the internal market and civil liberties was unequivocal. The draft law on the so-called “AI Act” is to be discussed in the plenum in mid-June, and negotiations with the council are to be completed by the end of the year. The topic has been on the table since February 2020, when the EU Commission first suggested regulations. The EU would be the first economic region in the world to set regulations for AI.

Now it’s about “differentiating between miracle tools and hell systems, for more civil rights and against total surveillance,” said Alexandra Geese, digital expert of the Greens/EFA parliamentary group yesterday. The committees’ proposal is aimed, for example, at restricting biometric surveillance (e.g. face and emotion recognition) in public spaces.

AI applications that represent “manipulative, exploitative and social control practices” are also to be banned. This includes, for example, algorithms that are used to create profiles and risk assessments in police work or to assess people’s creditworthiness. Such systems could deny people access to education, work or the social system, they say.

Austria’s state secretary for digitization, Florian Tursky, welcomed the agreement reached by the parliamentary committees. The VP politician referred to the Austrian AI authority, which is scheduled to go into operation in 2024. Dealing with AI is “crucial for prosperity in Austria,” said Tursky.