EU meets in Brussels to agree on legal framework

The negotiations promise to be tense. In Brussels, the institutions of the European Union – Parliament, Council and Commission – will try on Wednesday December 6 to agree on the draft regulation on artificial intelligence (AI). A major text while software like ChatGPT is growing and causing concern. And it is precisely these innovations that complicate negotiations.

While some are calling for the establishment of real barriers, countries like France are trying to slow down this regulation. There is no question of losing the race for generative artificial intelligence: this is the position of France, Germany or Italy. These three countries want to limit the barriers to innovation, as Emmanuel Macron recently recalled. “Beyond an industrial issue, AI indeed constitutes a civilizational challenge”, did he declare. “We need to ensure that the foundational models of artificial intelligence are not just fed with English.”

Bringing out European champions

If the United States has a head start with ChatGPT, Microsoft’s conversational robot or its competitor Bard, developed by Google, the Europeans also want to see champions emerge. In France, there is Mistral, in Germany, Aleph Alpha. These young companies are developing their own language model and campaigning against overly restrictive regulations. But European Commissioner Thierry Breton accuses them of defending their particular interest to the detriment of the general interest.

An agreement on Wednesday therefore remains uncertain and new negotiations could be necessary. Negotiators are meeting in Brussels from 3 p.m. and discussions are expected to continue late into the evening and even into the night. The legislative process on AI was launched in 2021 but progress around generative AI (as with ChatGPT) has changed the situation and forced the European Union to update its draft regulation.