“Artificial intelligence has no limits”: Italy, Germany and France block EU law

The first European law on artificial intelligence, aimed at countering the risks associated with the development of this technology, could hit a dead end. The “mistake” of the three big names of the European Union, namely the governments of Germany, France and Italy, for once clearly united around a common goal: to save the most advanced artificial intelligence models such as Gpt (better known by the ChatGpt chatbot) and Bard is subject to Brussels’ strict transparency and control rules.

Berlin, Paris and Rome said in a joint document that they did not agree with the compromise text towards which the Commission and the EU Parliament are increasingly moving. The regulation, proposed by the European executive in April 2021, aims to create a legal framework around artificial intelligence and avoid creating a wild west that can be intervened when things go wrong (or nearly so), as has happened with social media and online disinformation, for example. . Parliament has already voted in favor of strict rules covering all systems, including those deemed to be at highest risk. In particular, his proposal, which the EU Commission has also accepted, is to introduce transparency requirements for productive AI such as Gpt, which must be compulsorily adhered to, subject to a fine of millions.

These transparency requirements are intended to make it more clear to the public what moves within so-called “underlying models,” that is, the infrastructure from which AI systems considered potentially more dangerous originate. But for Germany, France and Italy, a strict regulatory network risks hindering the development of advanced models and slowing European companies’ competition with the rest of the world. Instead, the three main EU countries are proposing forms of self-regulation through the signing of commitments and codes of conduct.

It was understood that this was the orientation of Berlin, Paris and Rome, on the occasion of the tripartite summit held at the end of October between the German and French Economy Ministers Robert Habeck and Bruno Le Maire and the Italian Economy Minister. Adolfo Urso Enterprises. On this occasion, Habeck said, “Europe should not hide behind artificial intelligence. We have companies that are valid, strong and in many respects more capable than the technology giants of the USA and China.” Le Marie, on the other hand, rang the alarm bell: The French minister reported that “capital investments in artificial intelligence in 2022 are 50 billion in the USA, 10 billion in China and 5 billion in Europe.” This, he added, means that “Europe is investing ten times less than the United States and Beijing is investing half as much in the technology that is most important to our civilization.”

In order to pave the way for investments in Europe, the trio’s prescription is not to scare potential giants with very strict rules. However, some experts and most of the EU Parliament have a different view and state that artificial intelligence law cannot be strong with the weakest (i.e. the smallest systems) and weak with the strong. Canadian computer scientist Yoshua Bengio, a leading voice in the industry, warned that this would be “law of the jungle”. Moreover, even advocates of a lightweight approach, such as the United States, appear to have changed their minds.

The move by Germany, France and Italy could slow down EU regulations, if not completely derail them. Lawmakers had hoped to close the file by Dec. 6 so they could vote on it before the legislature closes in May 2024.

Continue reading today

Source: Today IT