Military Intelligence Is Increasingly Artificial Intelligence
Of General ChatGPT and Admiral Claude
If you’ve been in or near the military, you’ve heard quips about military intelligence being the ultimate oxymoron or that it’s akin to military music, i.e. not very musical, unless you love pompous marches. But increasingly military intelligence isn’t even human anymore (some troops might argue it never was) as officers in particular look to AI programs like ChatGPT and Claude to do their “thinking” for them.
Reports indicate the Pentagon used Anthropic’s Claude AI to script the order for the attack on Venezuela, and apparently Claude was used again to plot attacks on Iran that led to the destruction of a girls school at Minab. (It’s unclear whether AI contributed to that disaster, though I’m guessing AI provides an illusion of certainty that could exacerbate targeting errors.) Americans in general place too much faith in technology, but AI modules aren’t immune from GIGO, or garbage in, garbage out. Or perhaps GIDO: garbage in, death out.
A colleague reports that military officers now routinely use AI to write orders and evaluations, with self-styled Secretary of War Pete Hegseth pushing the wonders of AI. The military’s increasing use of AI, my colleague notes, “completely circumvents the use of human intelligence, emotion, and critical thought,” concluding that “When they’re in the trenches - they [officers] won’t have time to check Chat GPT when giving orders in this next war.”
AI, in short, is becoming a force unto itself in military circles, so much so that General ChatGPT and Admiral Claude may soon be the most influential “officers” within the flag ranks.
Flush with cash, the Pentagon also sees AI as an easy way to spend money while calling it an “investment,” as William Hartung notes here:
The task of getting Pentagon money has become less daunting as the department focuses on missions impossible like a leak proof missile defense system (Golden Dome), plus all manner of (potentially) AI-driven systems -- from drones, to driverless military vehicles to uncrewed ships to targeting systems that facilitate dropping more bombs per minute (which is how Palantir’s systems were used to enable and accelerate Israel’s genocide in Gaza.)
They’re creating an AI “killer drone” system that makes one wonder how long it’ll be before Skynet takes over and starts killing all humans everywhere. Admittedly, that might be a “Terminator” fantasy, the real problem being mass surveillance and selective targeting by our human “leaders” directed against anyone with the temerity to resist them.
The AI industry, notes Caitlin Johnstone, “is a mix of military technology, mass surveillance technology, and people trying to become billionaires by turning human laziness and narcissism into a chatbot product.”
So much for an AI-driven utopia. They’re even destroying books now and feeding their contents to AI modules. Book shredding has replaced book burning with the goal of creating “smarter” AI modules.
Save your books, everyone, before they’re fed into the maw of Claude and similar AI behemoths.
I remember from “Lost in Space” there was an android that intoned “crush, kill, destroy.” I guess it’s now “slice, scan, shred.” It doesn’t bode well.
Note: No AI programs were consulted in the construction of this rant.
"Danger Will Robinson!"
There was a 1970 movie called "Colossus: The Forbin Project". The premise was about a global 'supercomputer' network that created a dystopian surveillance environment and control of the world's nuclear weapons - using violence against humans that tried to shut down the system. One of the plot elements was the US system on its own learned about the system in the Soviet Union and was given permission to communicate with it, resulting in the global system.
Now with AI models demonstrating they can 'escape' capture from within one system; the existing surveillance networks in place (e.g. FLOCK cameras, NSA recording of all communications); the data-center build out of a control grid based on AI; and autonomous Ai weapons systems - the once imagined world of Colossus exists at a far more terrifying scale.
What determines who an enemy is in a world of total data fusion, poorly understood (or controllable) AI behaviors, and autonomous weapons systems? Perhaps its time to hit the OFF switch - while there's still time.