The European Parliament adopts a project to protect citizens from the dangers of AI

The European Parliament adopts a project to protect citizens from the dangers of AI

During the summer, we invite you to find articles distributed this year on the site. Today an article originally published on 15/06/2023.

Among the questions posed by artificial intelligence tools are those relating to public freedoms. Recent days have seen political momentum, especially in the West, regarding measures to limit. The European Parliament has just adopted a draft regulation to this effect, while the United Kingdom is trying to play its part and is planning a world summit on AI.

The European Parliament voted on June 14 in favor ofa regulation soberly titled "law on artificial intelligence" which had been prepared by the European Commission in 2021.

The project was adopted by an overwhelming majority, with 499 votes for, 28 against while 93 deputies abstained. It still has to be discussed with the Council of the European Union, in which the governments of the 27 Member States are represented. If agreed, it would be the first time in the world that a political organization has adopted such a text.

The process had been put on hold due to the emergence of new debates posed by ChatGPT and the image generator Midjourney, in particular the risk of information manipulation and copyright issues.

If the text is adopted, the users of these tools will have to declare whether already existing and protected documents have been used in the making of the new works, and the services offering these generative AIs will have to install a system making it possible to discriminate between the real ones and the fake ones. Pictures.

Recent images showing the Pope in a down jacket or Donald Trump on the run have gone around the world, and the use of images for politically manipulative purposes during election campaigns cannot be ruled out.

Social credit and mass surveillance

The project presents itself as a barrier to the principle of social credit and states that one of the areas in which "the use of AI systems deserves particular attention is access and the right to certain essential services and benefits, public and private, including health services and essential services, which include but are not limited to housing, electricity, heating/cooling and the internet, [in bold in the text, Editor's note] to enable citizens participate fully in society or improve their standard of living."

Parliamentarians also spoke out against the use of AI in the perspective of a mass surveillance society. The draft regulations oppose “real-time” remote biometric identification systems for law enforcement purposes, except for the prosecution of serious crimes and with judicial authorization.

It also provides for the prohibition of the untargeted collection of facial images from the Internet or CCTV footage for the purpose of creating databases that contradict human rights and respect for privacy. The text states as follows:

"The indiscriminate and untargeted capture of biometric data from social media or CCTV footage to create or expand facial recognition databases reinforces the perception of mass surveillance and can result in gross violations of fundamental rights , including the right to privacy. The use of AI systems with this destination in publicly accessible places should therefore be prohibited."

The vote comes a week after British Prime Minister Rishi Sunak announced a first global AI summit to be held in his country this fall.

While the United States and the European Union asked on May 31, during the fourth edition of the Council for Commerce and Technology, that a voluntary "code of conduct" be quickly put in place on the IA that the players in the sector would undertake to respect.

The United Kingdom, which has left the European Union, is keen not to stay out of the debate and hopes to host the headquarters of a possible global AI regulator.

Jean Sarpedon

Image credit: Shutterstock/Sinonimas

In the International section >

Recent news >