Speculative Essay on AI and Autonomous Systems Lacks Concrete Leads

the nature of physics brilliance. Boeing 747s lumbering across the Pacific towards San Francisco for decades faced the sweaty problem of cancelled landings as they circled above a fogged in airport, fuel running lower. The introduction of “autolanding” systems in the solved this for good. No big plane diverts from a misty field; it lands itself. AI offers the possibility of a kind of auto-land for our biggest physics puzzles, bringing them safely through a fog of data, theory and wrong ideas. But with this weird price: We may not fully understand why the answers are right. All around us Al-enabled systems will extend our ability to calculate and learn, to penetrate all sorts of foggy problems. They will sharpen our sadly dimming memories, keep us safe and even help us create. Just as those Al-enabled airplanes already make it impossible for pilots to fly into the ground, so computer wisdom may protect us from crashes of our own: Too much financial risk. Bad educational choices. (Poor music suggestions on a first date.) They will rely on their vast, instantly updated networks to tell us things we can’t see or would never notice in the first place: Don’t visit that office, everyone’s sick. They will use the ability to model thousands of possible outcomes of any choice to provide us with “feedforward” - an ability to learn from the future and not merely the past. Or, they will know to jam our brain full of the right chemicals at the right time: Here’s a Diplo track to put you in the mood to go forarun. You really need to exercise, Dave. Just as an age without connected devices will one day seem strangely antique, so will a world without the constant touch of AI. Recall Benjamin Franklin’s famous lament in the 1780s, that he’d sadly been “born too early” to enjoy the fruits of reason Starting to spill into his world as a result of the Scientific Revolution. Well, you and I (and scientists like Silk and Ellis) may have been “born too late” for an age of purely human cognition; the habits of connected thinking already inform our decisions and mark roads to new knowledge. The inevitability of Al reflects an inescapable logic at work now: We want faster better and smarter systems. We want to compress time. But the faster our world gets, the more it slips beyond a pace of human management. Al steps in. It makes the system function faster. Keeps itself safe. Us too.?62 Better-than-human Al inside these “representational” grids doesn’t vanish like it did in Maes’ lab. In fact, an honestly artificial intelligence is their nature of their strange essence. They will use it not simply to contemplate the world, to help us along, but also to confront what has never been seen, to see and then coldly manipulate any topology of power they can reach. Of course we'll still continue to think about the world; but the world, a wired and alive and cogitating cage, will think about us too.?63 262 Us too: Heinl, p. 53 263 Of course: Nigel Thrift and Shaun French, “The Automatic Production of Space”, Trans Inst Br Geogr NS 27 309-335 2002 192 HOUSE_OVERSIGHT_018424