The Pentagon's New AI Doctrine Is Watching Us All — doctrine…

The Department of Defense recently confirmed what defense contractors have been whispering about for months: a sweeping new artificial intelligence doctrine known as "Overwatch" is now operational across multiple branches of the U.S. military. The program, which pairs autonomous systems with human commanders in what officials call a "human-on-the-loop" arrangement, promises faster decisions, fewer casualties, and a decisive edge over adversaries. It also promises something else, though nobody in Washington wants to say it out loud: a permanent record of everything. At its core, Overwatch is a framework for continuous machine oversight. AI systems monitor communications, troop movements, supply chains, and even the emotional tenor of commanders' voices, flagging anomalies for review. In a battlefield context, that sounds efficient. In an American context, it sounds like the office of the future. Because the same contractors building Overwatch are selling its civilian siblings to police departments, hospitals, and Fortune 500 HR divisions. The doctrine doesn't stay in the desert. It migrates. It always migrates. Consider what this means for the average American worker. You clock in. A system logs your keystrokes, your tone in meetings, the micro-pauses before you answer questions. Your manager doesn't fire you; an algorithm flags you as "deviation risk" and the manager complies. You didn't do anything wrong. You simply moved in a way the machine didn't predict. This is the quiet catastrophe nobody is marching about. We are building a society where being watched is the default and being trusted is a privilege you earn through perfect legibility. The doctrine has a name now, but the feeling has been creeping into American life for a decade: the sense that you are always being graded, always being compared, always one anomalous data point away from being sorted out. The moral problem isn't the technology. It's the premise. Overwatch assumes that human judgment is the weak link, the thing to be corrected. But human judgment is also the thing that shows mercy. It's the manager who knows your kid was sick. It's the cop who lets you off with a warning. It's the commander who hesitates because something feels wrong. Every one of those moments looks like inefficiency to a machine. We are told this is about safety. It is always about safety. But safety for whom, and at what cost to the messy, forgiving, unpredictable business of being human? There's a deeper rot here too. Americans used to believe that privacy was a right, not a feature you toggle off. We used to believe that a person was more than the sum of their tracked behaviors. Overwatch, and its thousand civilian cousins, quietly repeals that belief. It replaces the citizen with the subject, the neighbor with the data point. The Pentagon will tell you the doctrine saves lives. Maybe it does. But doctrines don't stay in the buildings where they're born. They seep into the culture, into the way bosses manage, into the way schools discipline, into the way we look at each other across a parking lot, wondering who's recording. We didn't vote for this. We weren't asked. And that, more than any algorithm, is the tell. **The real danger isn't that machines will judge us. It's that we'll start believing they should.**