Former Google CEO Eric Schmidt has announced that he is backing a new non-profit AI lab based in Paris to carry out open research in the field of foundational AI models, such as those powering applications like ChatGPT — but warned that excessive regulation from Brussels could hinder European progress in AI.
The lab, called Kyutai, was announced by French billionaire Xavier Niel last September. In addition to Niel and Schmidt, it will also be funded by Rodolphe Saadé, the CEO of logistics multinational CMA CGM.
Commitments from the three patrons have already reached €300m and it is expected more private funds will follow.
Speaking at ai-Pulse, an event organised by Niel in Paris’s Station F startup campus on Friday, Schmidt said he backed Kyutai because he believes France can become an AI champion, as long as it can keep regulators in Brussels “under some level of control”.
“I’ve wanted to do this in France for a decade because it’s obvious that France should be a leader in this space,” said Schmidt.
“The maths and technical talent in France are world class. The problem is you’re stuck with Brussels.”
The EU AI Act
Schmidt — who is well-known for his criticism of EU regulators — pointed to the AI Act, which is currently being negotiated in Brussels, as an example of laws that might break innovation. The Act will impose new restrictions on businesses developing AI models based on the level of risk they pose to society.
His comments come at a sensitive time, with the AI Act recently hitting significant stumbling blocks as some EU countries — including France and Germany — push against rules placing extra requirements on foundational AI models.
European startups have expressed concern that the Act will significantly impact their ability to compete against US tech giants. Paris-based company Mistral, which develops foundational models, said that the text could “kill” the business.
Schmidt said that France should keep control of EU regulation to ensure that “the French can follow the same business path that the Americans are”.
“I think we will see leadership [in AI] in the UK, France and the US.”
Speaking after him, France’s digital minister Jean-Noël Barrot said that the French government is currently leading the discussions on the AI Act and working to convince other member states and the EU Parliament to drop the extra requirements on foundational models.
“We are adopting a radically pro-innovation approach to regulation,” said Barrot.
Kyutai: a new AI lab in Paris
Kyutai’s founding research team will be made of scientists that come from the ranks of Microsoft, Meta and DeepMind. The lab’s advisors include Meta’s chief AI scientist Yann LeCun, computer science researcher Yejin Choi and computer scientist Bernhard Schölkopf.
The event also saw French cloud provider Scaleway, which was founded by Niel, announce a series of measures to increase European AI startups’ access to the infrastructure they need to train their models.
They include a new partnership with chip giant Nvidia that will enable selected startups to benefit from priority access to GPUs — the much sought-after processors that are key to building AI tools.
It comes a month after Niel announced the purchase of Nvidia’s latest-generation supercomputer, specially designed to train foundational models, and which is available for startups to access through Scaleway.
Niel said at the time that creating sovereign AI models in Europe would require ramping up compute power on the continent.
High-profile AI startups like Poolside and Mistral, which build foundational models, have been using the supercomputer to train their models.