IA: European Parliament approves technology regulation – 03/13/2024 – Tech

The European Parliament approved this Wednesday (13) the world’s first regulatory framework to guide the use and development of artificial intelligence.

In preparation since 2021, the European Union’s AI Act divides technology into risk categories: low, medium, high and unacceptable, which would lead to a ban on the technology.

Citizen behavior scoring systems and manipulative AIs that offer incentives to maximize profits without compensation are prohibited. Facial recognition technologies in security were previously unacceptable, but were now considered high risk in the final text.

Chatbots and deepfakes are of limited risk and should signal the origin of the content on artificial intelligence platforms.

The rules come into effect at the end of the legislature in May, after final revisions and approval by the European Council.

Parliamentarians approved the proposal with a wide margin, 523 votes in favor, 46 against and 49 abstentions.

The EU had reached a political consensus on the text at the beginning of December, with positive signals from the ambassadors of the bloc’s 27 member countries. The vote was initially planned for April, according to AFP.

The text was changed after the launch of ChatGPT in 2022 to include so-called generative AIs, capable of generating text, images and sounds.

Germany and France, which are in the race to develop their own cutting-edge AI models, have raised doubts about the risk of legislation hindering technological advancement in the region.

In these negotiations, Germany and France wanted regulation to protect European companies specializing in AI.

French company Mistral.AI announced at the beginning of the month a model capable of competing with the paid version of ChatGPT and received a billion-dollar investment from Microsoft to make the service available on the Azure cloud platform.

European Parliament President Roberta Metsola called the act “pioneering” and said it would enable innovation while safeguarding fundamental rights.

“Artificial intelligence is already a very present part of our daily lives. Now, it will also be part of our legislation”, he wrote in a post on social media.

“Europe is now a global standard in AI”, published the European Commissioner for the Internal Market, Thierry Breton, on X (formerly Twitter).

The legislator responsible for negotiations in the European Parliament, Dragos Tudorache, praised the agreement, but stated that the biggest obstacle is implementation.