Table of Contents
French companies are eager to lead the way called France’s AI and curb the regulatory aspirations of the European Union. However, they face an unexpected adversary in the form of Thierry Breton, the European Commission’s self-proclaimed digital enforcer and a staunch advocate for France.
In late November, the EU’s internal market chief addressed a gathering of startups and tech investors in Marseille, where he criticized the French AI enthusiasts who are advocating for the weakening of the EU’s groundbreaking Artificial Intelligence Act.
France’s AI Public Interest Alignment
Breton highlighted Mistral, a France’s AI startup supported by renowned Silicon Valley investors, and grouped it together with the U.S. Big Tech companies that European Union officials typically enjoy criticizing.
During the event hosted by the French newspaper La Tribune, Breton acknowledged that Mistral is engaging in lobbying activities, which is a common practice. However, France’s AI emphasized that it is important to recognize that Mistral’s efforts are primarily aimed at protecting its own business interests rather than serving the broader public interest.
Breton revised his statements, expressing that Kyutai, an AI laboratory recently established with the support of two prominent French entrepreneurs, Xavier Niel and Rodolphe Saadé, along with former Google executive Eric Schmidt, is “a venture that serves the public good and embraces inclusivity. It is an endeavor that captivates and satisfies me.”
Breton’s statement stated the evidence, however, France’s AI community responded with confusion. “The perception was that he supported our cause,” reveals the public affairs manager of a French technology firm present at the Marseille event, speaking anonymously to freely address the matter. “Undoubtedly, Mistral serves private interests, but its contributions are equally accessible to the public, just like Kyutai’s.”
EU AI Regulations: France’s Stance on Foundation Models
Several French companies, with France’s AI Mistral being the most prominent, have recently contended that the regulation of advanced AI models would impede innovation in Europe. The French government, along with Germany and Italy, has shown support for these concerns and is resisting the proposed rules on “foundation models” put forth by the European Parliament.
The concept of a foundation model is a term used in Brussels to refer to the digital framework that supports robust and versatile AI systems like OpenAI’s ChatGPT. These models are widely recognized as the driving force behind the current AI frenzy. The Parliament aims to establish mandatory regulations regarding transparency and testing for the most advanced models, whereas governments suggest adopting voluntary codes of conduct.
The negotiations on the EU’s AI rulebook are drawing to a close, but the disagreement over foundation models could still jeopardize a final agreement. Breton’s strong statements are unlikely to win him much support in Paris. President Emmanuel Macron, a passionate advocate of technology who previously spoke enthusiastically about transforming France into a “startup nation,” has firmly backed French AI companies. He has even offered to host the third AI Safety Summit in November 2024. Last month, while addressing researchers, Macron emphasized the need for a global approach to France’s AI regulation in order to prevent hindrances for European companies.
Macron’s digital minister, Jean-Noël Barrot, has consistently criticized the regulatory plans of the European Parliament, highlighting France’s determination not to miss out on France’s AI gold rush as it did with the internet and social media. Mistral’s board includes Cédric O, who was Barrot’s predecessor. Breton’s comments are likely to cause further annoyance for another reason. France’s strategy has long relied on attracting young AI developers away from major U.S. companies.
Franco-U.S. Tech Collaboration and Concerns Over AI Regulations
The Mistral itself is a prime example, with its three co-founders coming from Google and Meta. However, in recent months, France has become increasingly open to collaborating with U.S. companies in order to overcome technological barriers and develop its own ecosystem.
During the recent incident involving Sam Altman’s removal as CEO of OpenAI, the invitation extended by Barrot to Altman to relocate to France’s AI was prominently showcased. This development was partially orchestrated by Yann LeCun, a renowned French researcher and Meta’s Chief AI Scientist.
LeCun played a pivotal role in initiating Meta’s Paris research lab, FAIR, as early as 2015. According to an anonymous startup lobbyist, who requested anonymity due to lack of authorization to speak publicly, France is fortunate to have LeCun. He is considered one of the pioneers of France’s AI, making him a revered figure within the industry.
It should be noted that Breton is not the sole advocate for France’s AI regulation. Mark Brakel, the policy lead at the global AI-focused think tank, the Future of Life Institute, agrees with Breton’s stance and warns that lobbying efforts by Mistral and other tech companies could potentially prioritize the profits of a few at the expense of the benefits for many. Additionally, the French cultural sector has raised significant concerns about the potential sacrifice of intellectual property in favor of technological advancement.
Intellectual Property Concerns Spark Debate in the French AI Ecosystem
Pascal Rogard, the president of the authors’ society SACD, stated to Le Monde that France, the birthplace of copyright, has failed to protect intellectual property for the first time. In response to this, around 80 unions and organizations from the cultural sector wrote a letter to the government expressing their disagreement with the French position on the AI Act.
As these concerns continue to rise, the French AI ecosystem is prepared to retaliate. Just recently, Yann Le Cun issued a strong warning during an event organized by platform regulator Arcom regarding the consequences of enforcing copyright for generative AI models. “I want to make it clear that if copyright is enforced, the AI industry will come to a halt. It simply cannot function without it,” he emphasized.
