The EU AI law was passed, but Mistral continued |  The AI ​​beat

Unveiling the Triumph of Mistral: A Game-Changer in AI Landscape

After a three-day negotiating marathon, the EU AI Act was finally passed this weekend with Carme Artigas, Spain’s Foreign Minister for Digitalization and Artificial Intelligence, calling it “a historic achievement and a huge milestone on the way to the future!”

Among other agreements, the EU AI law requires systems labeled as “general purpose high-impact AI models” to meet transparency standards,

while systems classified as “high risk” are subject to additional requirements, including risk management, monitoring serious incidents, evaluating the AI ​​model, and implementing red teaming strategies.

In other words, the EU AI Act was huge, champagne-popping AI news – about a minute long.

It lost some momentum in the media when it became clear that the law is a provisional agreement, meaning it could take months for the bill to receive its final wording.

Moreover, it will not be enforced anytime soon: the AI ​​law will not come into effect for up to two years after final approval from European legislators.

Um… two years? That’s the equivalent of two centuries in the AI ​​hype cycle.

The EU AI Act was also quickly pulled from the headlines by one of Europe’s own AI stars.

Paris-based open-source model startup Mistral released, in the equivalent of a microphone drop, a new LLM on Friday with nothing but a torrent link.

Yesterday, the company, which was founded just seven months ago by Meta and Google researchers, kept the party going by announcing a $415 million fundraising from investors including Andreessen Horowitz, Lightspeed Venture Partners, Nvidia, and Salesforce – with the little 22 year old was appreciated. person team at approximately $2 billion.

Mistral had led the charge against regulating foundation models

It is notable that Mistral was at the forefront of negotiations on the EU AI Act a month ago and lobbied against the European Parliament’s proposal of a layered approach to regulating generative AI that ensured that foundation models had guardrails.

Many pointed to the fact that Mistral had deep ties to the government – ​​Cedric O, the former French Secretary of State for Digital Affairs, was the one who led Mistral’s lobbying efforts in the EU AI Act.

Back in October, O had said in an interview that the EU AI law could be “a zero-or-one regulation” for European generative AI startups like Mistral.

“Depending on [final shape of the] AI Act, it could kill Mistral, it could allow Mistral to grow, and there’s a lot of gray area in between,” he said. “It’s crucial.”

Ultimately, Mistral appears to have succeeded in its lobbying efforts: the agreed EU AI law provides broad exemptions to open-source models (although these do not apply to open-source models deemed to pose a systemic risk).

The major victories of the past few days make it clear that the EU AI Act is certainly not the end of the story when it comes to AI regulation in the European Union.

An article in TechCrunch made a good point about the debate going on there at the moment: Union: “Mistral’s fortunes are in many ways a microcosm of the battle for AI sovereignty.

The European Union (EU) wants to avoid being left behind in yet another technological leap, while at the same time imposing regulations to guide the development of technology.

As the German Vice Chancellor and Minister of Economic Affairs Robert Habeck recently quoted saying: “The idea of ​​self-sovereignty in the AI ​​sector is extremely important. But if Europe has the best regulations, but no European companies, we have not gained much.”

AI regulation will continue to catch up

Be that as it may, it is clear that AI regulation – whether in the EU, US, or elsewhere – will have a long way to catch up.

The report that Mistral’s new Mixtral 8x7B model eclipsed the performance of OpenAI’s GPT 3.5 came as thousands of AI researchers, including those from all the top AI labs, gathered at NeurIPS (Conference and Workshop on Neural Information Processing Systems) in New Orleans.

Those researchers, who have just completed a mind-blowing year in generative AI, are also ready to party – at least once they finish presenting their papers.

They will raise their glasses in a technological toast to even faster AI development in 2024, as policymakers around the world try to figure out how to rein it in without turning back the dial on innovation.

If the grand celebration of Mistral is any indication, regulators have one big need right now. A need for speed.

VentureBeat’s mission should be a digital town square where technical decision makers can learn about transformative business technology and transact. Discover our Briefings.

2 thoughts on “Unveiling the Triumph of Mistral: A Game-Changer in AI Landscape”

  1. We’re a group of volunteers and starting a new scheme in our community. Your site provided us with helpful info to work on. You have performed an impressive job and our whole neighborhood might be grateful to you.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top