Autonomous targeting without human control and untargeted profiling formed the core red lines of a new government contract framework. Developed at Google, the system balances operational security with auditing requirements. It specifically avoids legal loopholes that previously weakened Anthropic's policies. This provides a concrete blueprint for negotiating ethical constraints in defense AI procurement.