MEPs on Wednesday adopted rules to regulate artificial intelligence (AI) systems such as ChatGPT, the first law of its kind in the world, France Presse reported.

Thierry Breton, the European commissioner in charge of the file, welcomed the X parliament’s “overwhelming support” (523 votes to 46) for the text, which sets out “the world’s first binding and comprehensive rules for reliable AI”.

The draft law was presented by the European Commission in April 2021. The appearance at the end of 2022 of ChatGPT (launched by the American company OpenAI), capable of writing essays, poems or translations in a matter of seconds, made the project gain popularity. new value

ChatGPT revealed the enormous potential of AI, as well as the risks it carries. The spread of fake photos and videos highlighted the danger of manipulating public opinion.

“This is just the beginning”

“Today is a historic day in our long journey to regulate artificial intelligence,” co-sponsor of the text Brando Benifay (Social Democrat) told the media on Wednesday morning.

With this text, “we managed to find a very delicate balance between the interest in innovation and the interest in protection,” added co-rapporteur Dragos Tudorake (Renew, centrists and liberals).

However, this legislation “is just the beginning,” he said, noting that artificial intelligence continues to develop rapidly. “We will have to be very attentive to these technological developments in the future and be ready to respond to new challenges that may arise,” Dragos Tudorake warned.

What the legislation on artificial intelligence provides

The legislation provides for a two-tiered approach. AI models for “public use” will have to comply with transparency obligations and European copyright rules.

Systems that are considered “high risk” and are used, for example, in critical infrastructure, education, human resources and law enforcement will be subject to stricter requirements.

For example, they will have to include a mandatory assessment of the impact on fundamental rights. Artificially created images, text or videos (deep fakes) must be clearly identified as such.

The text also bans citizen ratings and mass surveillance systems similar to those used in China, as well as remote biometric identification of people in public places.

However, with regard to this last point, Member States have been granted derogations for certain law enforcement missions, such as the prevention of a terrorist threat or the specific search for victims.

European AI Office

European legislation will be provided with means of monitoring and sanctions, and a European Office for Artificial Intelligence will be established within the framework of the European Commission. It will be able to impose fines of between 7.5 and 35 million euros, depending on the crime and the size of the company.

“We regulate as little as possible, but as much as necessary,” European Commissioner Thierry Breton wrote on X.

But the tech world is wary.

“Many of these new rules remain unclear and could slow down the development and implementation of innovative programs,” said Boniface de Champry, head of European lobby group CCIA.

For their part, Observatoire des Multinationales (France), Corporate Europe Observatory (Belgium) and LobbyControl (Germany) fear that lobby groups will undermine the implementation of AI rules.

“Many details of the AI ​​law remain open and require clarification (…), such as standards, thresholds or transparency obligations. The composition of the advisory board of the new European artificial intelligence agency also remains unclear,” they warned in a joint statement.

The 27 EU member states are expected to approve the text in April, and the law will be published in the Official Journal of the EU in May or June.