The Pentagon's AI Problem Is Degraded Thinking, Not Rogue Drones

The standard most cited as the safeguard for U.S. autonomous weapons policy was designed to prevent one specific failure mode: a machine making a lethal decision with no human authorization. It was not designed to assess whether the human authorizing that decision retained the analytical capacity to make it. exposes the gap between what the standard promises and what it delivers — a human who clicks 'confirm' at the end of an AI-generated recommendation chain is present in the…