Green MEPs ask EU to look into ethics of Mistral AI’s Microsoft partnership

Green MEPs ask EU to look into ethics of Mistral AI’s Microsoft partnership

Microsoft’s partnership with French start-up Mistral AI is stirring a political fallout in the EU with members of the European Parliament (MEPs) voicing concerns about Big Tech taking over and questions arising over who knew what when the EU AI Act was being negotiated.

The European Commission said on Tuesday it would look into the “strategic partnership” announced a day earlier, which sees a €15 million investment in Mistral AI and Microsoft owning a “small” stake in the company.

Mistral’s AI models will also be available via Microsoft’s Azure cloud computing platform.

But Green MEPs also want the Commission to look at the “ethical side of the deal,” not just competition.

This week the group sent a letter to the Commission asking it to look into possible conflicts of interest and transparency issues about Microsoft’s and Mistral's lobbying.

“Microsoft has been lobbying intensely since the start of negotiations,” Kim van Sparrentak, a Dutch MEP from the Greens who worked closely on the AI Act, told Euronews Next.

“Mistral was really lobbying the European Council,” she said, adding that she is not criticising the business decision for a partnership.

Negotiations for the EU AI Act began in November last year and lobbying was at play.

France, among a couple of other EU members, fought to water down the rules to protect open source companies such as Mistral, arguing that European companies would turn to the US and China for investment.

“No one can say if Mistral was in talks with Microsoft back in November,” said Michelangelo Baracchi Bonvicini, president, of the AI4People Institute, which was founded to shape the debate on AI ethics in the EU in 2018.

“But we can say for sure that it is unlikely to have a partnership with such a Big Tech company in just a month. It is not even possible to speak to Microsoft’s chief technology officer with two weeks' notice,” he told Euronews Next.

Mistral has turned down Euronews Next’s interview requests but the company’s co-founder and chief executive Arthur Mensch said on social media there were “creative interpretations” of the deal and that it is “an independent European company with global ambitions”.

Another question for MEPs is whether the French government knew anything about the deal before the Act’s negotiations.

Van Sparrentak said there are “many levels” to the story as France’s former digital secretary of state, Cedric O, sits on Mistral’s board and has a close relationship with the French government.

Meanwhile, seeing as Microsoft invested heavily in ChatGPT-maker OpenAI, van Sparrentak argued that a possible OpenAI office in Europe would be an opportunity for France to house it in Paris.

“It seems odd if no one in the French Parliament had [any] idea about the deal but did everything to change the Act,” she said.

But a spokesperson at the French finance ministry told Euronews Next it did not know about the partnership before the announcement and saw the deal as a source of pride.

The spokesperson also said France did not try to water down the EU AI Act for Mistral in particular.

Microsoft’s monopoly

Microsoft has invested billions in AI and integrated OpenAI’s models into its products, such as Office 365.

Van Sparrentak said she is “concerned about Microsoft taking over the market,” and said she was disappointed there is no investment in Europe’s cloud infrastructure.

Others are more worried about ethics than European sovereignty.

“The future will not be important whether AI is American or European but [rather] if it is good AI - which respects the rules - or bad AI - which does not respect the rules,” said Baracchi Bonvicini.

Euronews has contacted Microsoft for comment but did not receive a reply at the time of publication.