Photo: Jean-François Badias Associated Press MEPs adopted the Artificial Intelligence Act on Wednesday in the European Parliament in Strasbourg.
Pauline Froissart – Agence France-Presse and Raziye Akkoc – Agence France-Presse respectively in Strasbourg and Brussels
3:50 p.m.
- Europe
European MPs adopted rules on Wednesday to regulate artificial intelligence (AI) systems like ChatGPT, a unique approach in the world.
The European Commissioner responsible for the file, Thierry Breton, welcomed on
“This will benefit Europe’s formidable talent pool. And will establish a model for trustworthy AI throughout the world”, underlined the President of the European Commission, Ursula von der Leyen, referring to a “pioneering” law.
This bill was presented by the European Commission in April 2021. The appearance at the end of 2022 of ChatGPT, from the young Californian startup OpenAI, capable of writing dissertations, poems or translations in a few seconds, gave it a new dimension.
This conversational agent revealed the enormous potential of AI, particularly generative, but also its risks. The dissemination of false photos or videos, larger than life, has thus alerted to the danger of manipulation of opinion.
“Let the beginning”
French President Emmanuel Macron welcomed the vote, calling it “a first in the world, essential to protect everyone’s rights and data security while supporting innovation.” “It’s Europe that’s doing it! » emphasized the head of state on IA] than other » countries.
With this text, “we have managed to find a very fine balance between the interest to innovate and the interest to protect,” declared co-rapporteur Dragoş Tudorache (Renew, centrists and liberals).
However, this law “is only the beginning,” he noted, emphasizing that artificial intelligence continues to evolve rapidly.
The legislative document provides for a two-tiered approach. “General purpose” AI models will have to comply with transparency obligations as well as European copyright rules.
As for systems considered “high risk” — used for example in critical infrastructure, education, human resources, law enforcement — they will be subject to stricter requirements.
They must, for example, provide for the establishment of a mandatory analysis of their effects on fundamental rights.
Artificially generated images, texts and videos (which may in particular lead to hyperfaking) must be clearly designated as such.
The text prohibits citizen rating or mass surveillance systems used in China, or even remote biometric identification of people in public places.
On this last point, States have however obtained exemptions for certain law enforcement missions, such as the prevention of a terrorist threat and the targeted search for victims.
The law will be accompanied by means of surveillance and sanction with the creation, within the European Commission, of a European AI office. It can impose fines ranging from 7.5 to 35 million euros (11 to 51.5 million Canadian dollars), depending on the offense and the size of the company.
“We regulate as little as possible, but as much as necessary,” summarized Mr. Breton.
“Gaps, restrictions and exceptions”
The European Consumer Bureau, however, indicated that “the legislation should have gone further to protect consumers”.
“The final text is full of loopholes, restrictions and exceptions, meaning it will not protect people or their rights from some of the most dangerous uses of AI,” the group echoed. digital rights advocacy Access Now.
For his part, Markus J. Beyrer, director general of BusinessEurope, the voice of European employers, considered that this was a “pivotal moment for the development of AI in Europe “.
“But successful implementation will be key to reassuring investors, AI developers and citizens that trustworthy AI is here for good,” he said. he continued.
The Multinational Observatory (France), Corporate Europe Observatory (Belgium) and LobbyControl (Germany) fear that lobbies will weaken this implementation.
“Many details of the AI law remain open and need to be clarified […], for example regarding standards, thresholds or transparency obligations. The composition of the advisory board of the new European AI agency also remains unclear,” they warned in a joint statement.
The 27 states of the European Union are expected to approve the text in April before the law is published in the Official Journal in May or in June.