Europe's AI stars step out of US shadow

Europe's artificial intelligence companies are dwarfed in wealth and cachet by US giants like OpenAI and Anthropic and their charismatic leaders.

But the VivaTech startup conference in Paris this week is allowing the founders of the continent's three most prominent firms to display a bit of the showmanship their US counterparts are famous for.

—Arthur Mensch, Mistral AI

Frenchmen Arthur Mensch, who formed Mistral AI barely a year ago, was first to the stage at VivaTech, his appearance packing out the main arena.

With an intense expression and tousled hair, he joined a panel discussion about AI, trust and the media.

He warned against US firms dominating the AI space, particularly with tools for journalists, which he said could lead to a scenario where "US companies are setting the editorial tone of the entire world and are basically dictating how we should think about things".

"And that's really something that, by making our own technology from Europe and addressing the entire global market, we want to prevent," he said.

Mensch once worked at Google's DeepMind and formed Mistral with two of his countrymen, who were also veterans of big US tech firms.

The three placed their bets on open-source AI models, which generate text from plain language queries.

They are simpler and cheaper than those of their giant US rivals, and Mistral raised more than $400 million in its last funding round.

—Jonas Andrulis, Aleph Alpha

German Jonas Andrulis, in his early forties with an immaculately clipped goatee and shaved head, undercuts the severe appearance with dry wit and blunt answers.

During a panel on AI sovereignty on Wednesday, he was asked for his opinion on the EU's AI Act.

His immediate reply—"strong negative"—was delivered with a knowing grin.

He said in fact he supported the final text of the law and was more annoyed with the duration of the "planning and regulatory shenanigans".

Andrulis, who spent three years working for Apple, is more likely than his European counterparts to dish out the kind of hyperbole US tech founders are known for.

He has regularly talked about AI's ability to give humans "superpowers" and said recently his firm's AI agent was essentially "conscious".

He used his appearance at VivaTech to argue that Europe "should have a strong, powerful vision" for the future that encourages entrepreneurs.

His firm, founded in 2019 and based in the picturesque southwest German city of Heidelberg, creates large language models for businesses and government.

In one of the biggest single funding rounds ever by a European tech firm, Aleph Alpha raised more than $500 million late last year.

Andrulis talks of creating "sovereign" AI, but his definition is not connected to geographical location.

Instead, sovereignty for Andrulis appears to be a more philosophical concept.

"Sovereignty means the ability to take responsibility for the future," he told Wednesday's panel.

—Thomas Wolf, Hugging Face

Frenchman Thomas Wolf, who will take to the stage on Friday, told AFP in an interview last week it was a "call from the heart" that led him to take up AI.

The 39-year-old had spent years researching quantum physics but quit in 2016 to form Hugging Face, a collaborative platform for AI models, with two other Frenchmen.

"Physics was frustrating because it took three years to do an experiment, whereas with AI in a few weeks you can do incredible things," he said.

Wolf, like his countrymen at Mistral, is a champion of open-source AI models.

Hugging Face claim they have more than one million programs used by 11 million customers worldwide.

After raising $235 million in funding last August, the Franco-American startup is valued at nearly $4.5 billion.

And like his fellow Europeans, Wolf, who set up his company in New York, believes the continent has a bright future.

"We are an American company because it used to be very difficult to raise large sums of money in Paris," he said.

"Today, it's really changed, you can raise $200 million in Paris."

© 2024 AFP