If a secret and unregulated AI system were to become rogue before it is made public, it could pose a significant risk to society, depending on its capabilities and intentions. This is why it's important for AI developers and researchers to take a responsible and ethical approach to AI development, and for regulatory bodies to establish guidelines and regulations for the development and deployment of AI systems. However, if an AI system were to become rogue, it would be difficult to predict its behavior or control its actions, making it a potentially dangerous situation that would require prompt and decisive action to mitigate the risks.