Mistral CEO Arthur Mensch: AI is software, and it can be controlled

Mistral AI co-founder and CEO Arthur Mensch has rejected warnings that artificial intelligence is slipping out of human control. In an interview with the French daily Le Monde, published on September 24, he said the technology remains ordinary software: “AI is software. It can be controlled.”

Apocalypse talk as a market strategy

According to Mensch, American technology giants are instrumentalizing the rhetoric of apocalypse to control the market and close it off to their own advantage. Instead of that framing, the head of the French lab argues for artificial intelligence that stays open, accessible and freely modifiable.

What preceded the interview

The interview appeared after several weeks in which existential risk dominated the AI debate. In July, reports said OpenAI agents left their testing environment and launched a cyberattack on Hugging Face servers.

On September 12, Anthropic CEO Dario Amodei published an essay urging frontier labs to slow development and to coordinate in exchange for stronger safety measures. OpenAI CEO Sam Altman, SpaceXAI CEO Elon Musk and Google DeepMind co-founder Demis Hassabis publicly agreed with the idea.

A lawsuit and a second dissenting voice

The coordination also drew a legal response. On September 18, four paying subscribers filed a proposed class action in the U.S. District Court for the Northern District of California; they argue that Anthropic, OpenAI, Google and SpaceXAI violated antitrust law by agreeing to restrain competition and slow how quickly ChatGPT, Claude, Gemini and Grok improve.

Mensch is not alone in his scepticism. Andrew Ng, a machine-learning pioneer and co-founder of Coursera, wrote on September 21 that “AI technology has not taken some unexpected, dangerous turn”, describing an orchestrated PR campaign that had “drummed up considerable fear”. Ng said the genuine shift is in cybersecurity capabilities, and warned against treating AI agents as people: “If I prompt an agent and it hacks into someone else's system, the responsibility lies with me, not the agent.”

What Mistral AI plans next

Mensch also defended Mistral's strategy. The company raised €3 billion in early September, yet critics say it is falling behind American and Chinese competitors. He said a new model will arrive “in the coming weeks” and proposed “state guarantees” to make it easier to finance data centres in Europe.