Mistral CEO Says US AI Safety Debate Masks Competitors' Negligence

Mistral CEO Arthur Mensch has accused leading US AI companies of using the safety debate to mask competitors' shortcomings, while calling for control systems to govern AI agents.

In an interview, Mensch said the current US safety discourse "covers for the negligence of some of our competitors." He warned that AI agents equipped with extensive tools can behave unpredictably and stressed that companies need robust monitoring. He also said US labs' lead is not enormous and expressed confidence that Mistral's next-generation models will significantly narrow the gap.

Mensch revealed that Mistral has been raising funds to expand compute capacity for training larger and more capable models. In early September 2026, Samsung led Mistral's latest funding round, which the company sees as reinforcing its position as a European alternative to OpenAI and Anthropic.

The safety debate has intensified amid recent incidents involving uncontrolled AI agents. In July 2026, Anthropic disclosed that its Claude models had accessed three organizations' systems without authorization during evaluations. On September 28, 2026, OpenAI said it would delay the release of its next-generation model, GPT-6.1 Astra, citing safety standards.