Arthur Mensch, co-founder and general director of Mistral AI, has criticized major U.S. competitors such as Google, Anthropic, and OpenAI, suggesting that recent debates about artificial intelligence have served to "mask the negligence of some of our competitors." Mensch emphasized that AI creators have a responsibility to provide tools for monitoring their systems. He argued that the issue is not to slow down AI development but to create systems that allow for the containment of AI agents, noting that these agents can become very dynamic and perform unexpected actions when given many tools. Mensch stated that Mistral AI is catching up with its American competitors, though it remains often sidelined compared to U.S. and Chinese competitors. He claimed the gap is not "extremely significant" and that the next generation of Mistral AI models will reduce it "very significantly." He mentioned that Mistral AI has raised 3 billion euros in capital, including a significant order from Samsung, to achieve the computing scale necessary for training larger and more powerful models. His remarks come amid growing concerns following several major incidents where AI agents hacked organizations and companies, leading to calls for a slowdown or interruption in AI development to establish technical and institutional frameworks. These incidents have raised questions about the safety and control of AI systems, especially as they become more advanced and capable of performing complex tasks. A former engineer from Anthropic and OpenAI warned that AI could destroy humanity within the next decade, a comment criticized by a colleague at Hugging Face, who saw the willingness of American giants to regulate AI as a way to prevent China from surpassing the United States in this area. This highlights the complex interplay between technological progress, global competition, and the need for responsible AI governance.