Mistral CEO Arthur Mensch is advocating for open-source AI, warning companies that relying on closed models can give labs a window into their business processes, potentially allowing them to exploit that information. Mensch advises companies to store their data in open systems and build their own training models to maintain control. This approach may seem daunting, but it’s crucial for companies to have agency over their AI growth. Why it matters: as AI becomes increasingly integral to business operations, the debate over open-source versus proprietary models has significant implications for data ownership and security in the broader AI landscape.