A new Cyberwarfare competition will pit offensive and defensive AI algorithms in battle against each other. Kaggle.com is providing the platform for the data science competition where algorithms test their metal confusing and thwarting each other. The practice combat is intended to illuminate how best to strengthen machine learning to defend itself against attacks. Is Cyberwar the new NFL? This is no game. The security and viability of machine learning depends on our ability to harden the muscles of these systems to both fool enemy deep neural networks and create algorithms for deep neural networks that cannot be fooled. Deep learning depends on huge, interconnected webs of mathematically simulated neurons. As machine learning and deep learning continue to establish themselves as essential tools in industries and governments, creating an effective “immune system” to prevent and react to attackers is essential. The contest will have three components: confusing a machine-learning system to keep it from working; forcing a system to classify something incorrectly; and developing robust defenses. The results will be presented at a major AI conference later in 2017.