California Draws a Hard Line on AI: No State Contracts Without Labels, Safeguards and Accountability

California has moved to set one of the clearest state-level boundaries yet for artificial intelligence in government, tying public contracts to privacy standards, security rules, and visible labeling of machine-generated content. The message is blunt: if companies want state business, they will need to show not just innovation, but controls. That turns procurement into leverage, and leverage into policy. The most consequential piece may be the watermarking requirement. As synthetic text, images, and video become harder to distinguish from human-made material, even a state-focused labeling rule could ripple outward into national practice. For the tech industry, this is a signal that the era of selling AI on speed alone is fading; governments are starting to ask what happens when these tools fail, mislead, or scale carelessly.