The governments of Germany, France and Italy have seen the development potential of artificial intelligence, so they have recently changed their attitude and issued a joint agreement stating that Europe needs an AI regulatory framework that promotes innovation and competition. This is diametrically opposed to the strict regulatory attitude previously conveyed by the EU. The EU wants to fully regulate and impose tight restrictions on the most powerful AI models. However, Germany, Italy and France obviously value the possibility and competitiveness of self-developed large models, and would rather sacrifice some of their regulatory powers for this purpose.
Germany, Italy and France pointed out in the joint document that European participants should strive to stand out in the global artificial intelligence competition and convey Europe's voice and values. AI regulation can be self-regulated through company commitments and codes of conduct.
A member of the European Parliament's negotiating team responded by calling it a declaration of war.
The big brothers were the first to rebel.
The EU’s ambitions in artificial intelligence are both secretive and overt. Although the EU has always claimed that it will cooperate with the international community to develop a global AI regulatory system, its goal is to become the top artificial intelligence regulatory agency in the Western world, and most of the time it is a maverick.
At this time, Germany, Italy and France hope to integrate with international standards, but this represents a betrayal of the EU. It is said that after hearing that the three countries had launched an association agreement, the relevant negotiators of the European Parliament chose to withdraw from the meeting of government representatives of the European Council and European Commission officials, and legal negotiations at the EU level have stalled.
Negotiators have until December 6 to complete the talks, but hope currently looks slim. Considering that the European Parliament will hold re-election elections in June 2024, it seems difficult for the EU-level AI regulatory framework to pass legislation before this point in time.
In addition to Germany, Italy and France, other EU countries, especially Spain, which holds the rotating presidency of the Council, have been emphasizing the expansion of the regulatory scope of the Artificial Intelligence Act and requiring that those basic models be covered. But with Germany, Italy and France forming an alliance, it will be difficult for other countries to win on this point.
Harry Borovick, general counsel of AI company Luminance, pointed out that this move by Europe's three largest economies may plunge the EU's regulatory discussions into chaos. The joint document is indeed good for business, but the effect is not great because it is based on the voluntary basis of the company. But the document is certain to divide the EU and delay discussions.
How OpenAI’s drama shocked Europe
A highly controversial point is that Germany, Italy and France hope to relax regulations on basic models and encourage industry autonomy, but this idea is considered by some to be as ridiculous as "the emperor's new clothes".
Canadian computer scientist Yoshua Bengio pointed out that it is crazy to ignore the management of basic models and will make the EU's artificial intelligence bill like the law of the jungle. Benign artificial intelligence systems are strictly regulated in the EU, but models that are dangerous or potentially dangerous go unpunished.
At this time, the OpenAI palace drama far across the ocean also shocked the European Union. Brando Benifei, one of the two negotiators in the European Parliament, pointed out that the OpenAI situation shows that governments cannot rely on voluntary agreements brokered by industry leaders.
Dutch Digital Minister Alexandra van Huffelen also added that OpenAI’s internal struggles highlighted the shortcomings of the AI industry’s lack of transparency and over-reliance on a few large companies, which emphasized the need for regulation.