DARPA awards 'Deep Green' battle-computer cash
General-ware to direct human pawns on future battlefields
Arms globocorp BAE Systems has won a contract to develop a thinking battle computer able to "help" human commanders directing US forces in the wars of tomorrow.
While the "Deep Green" project is billed more as a means of automating tiresome staff officers out of a job, its name suggests otherwise. Deep Blue, as everyone knows, was the first computer to surpass the world's top-ranked human grandmaster in the mock combat of chess. It seems safe to say that Deep Green is intended not merely to assist human generals but ultimately to become far more adept than them in the art of battlefield command.
For now, the Pentagon announced on Friday that BAE Systems National Security Solutions in Massachusetts was awarded a $5m-plus contract last week "for the development of a battle command decision and support system" in connexion with the Deep Green programme. People often think of BAE as a British company, but in fact it nowadays has more US employees than UK ones as it shifts base across the Atlantic.
Given the nature of the project, the identity of the Pentagon funding body won't come as a surprise. It is DARPA, of course, the maverick military research bureau whose boffins collectively seem to paraphrase Sir Isaac Newton and say "if we are to see further than others, this will clearly mean standing on the shoulders of gigantic, tremendously powerful robotic dinosaurs able to shoot death rays out of their eyes".
For now, the Deep Green kit is intended merely to reduce the number of brown-nosing staff officers in the retinue of American brigade commanders. This means that even once the digital Napoleons throw off the yoke of their human bosses, any sneaky plans for machine genocide against humanity will still be hampered. Most of the units under their command will be composed of humans - for now, anyway. (Fortunately, the US Army's recent plans to purchase a robot tank have been temporarily binned.)
Even so, there are already well-advanced programmes intended to field heavily armed droid kill-choppers - not to mention nasty little flying-dustbin pocket Daleks, which could be troublesome even inside buildings - or even robot ornithopter spy bats featuring gargoyle stealth mode. And these are just the army machines.
If the AI command software manages to get into the air force or navy hierarchies and gain control of their fearsome fleets of robot bombers, there won't be much left for humanity to do. Some kind of technically implausible deal involving the machines harvesting our bodies for power would probably be our only hope of survival. Luckily, DARPA is also working on a Matrix-style war simulation network which could easily become a fake world in which our disenfranchised brains could live a meaningless, empty, tightly regulated life.
Assuming that the humdrum scenario of humanity's extinction or enslavement at the hands of its erstwhile computer servants doesn't come to pass, there's a more zeitgeisty theme to worry about. BAE has lately been accused in barely veiled terms by US defence officials of pirating top-secret American war-tech out of its US operations, perhaps for worldwide sale.
The Defense Inspector-General couldn't really back that up. Even so, software is notoriously difficult to keep closely held. It might be only a matter of time before the invincible US forces, directed by software packages infinitely superior to any human general, found themselves up against enemy forces under the control of pirated warez every bit as good. ®
Sponsored: DevOps and continuous delivery