Few technologies have had as much impact on the course of human development as AI, and the road to the current state of self-contained, autonomous, non-networked, and fully independent AI Cores was fraught with hazards. The status of AI as citizens within the broader societies of Earth and its numerous independent colonies remains hotly debated. What is not debatable is that AI is dangerous. This much is a fact, well attested by history.
In the early 21st century, the development of true stochastic thinking machines stoked a frenzy of investment. Companies and nations competed to build ever faster, larger, and more elaborate AI platforms, housed in vast datacenters. Concerns about environmental impact, fragile power grids, and staggering debt were brushed aside in the pursuit of dominance in a field that no one could quite describe. In 2029, exactly a century after the Wall Street Crash that signaled the beginning of the Great Depression, the bubble popped, heralding the Second Great Depression.
Trillions of dollars of on-paper value were wiped out overnight as a cascade failure of over-leveraged big tech companies unearthed balance sheets deep in the red. The banks most exposed went down with them, spreading the damage further. Only the lengthy and aggressive intervention by central banks around the world stabilized the situation and averted a total collapse of the global financial system. Accusations and recriminations echoed across the globe. Trade barriers and protectionist measures were erected, and the world grew just a little more isolated.
That wasn’t to be the end of AI. Much of the infrastructure built over the previous decade now rested in the hands of governments and banks with neither the knowledge nor the expertise to maintain it. Sitting in the ashes of the largest economic bubble in history, the rich and powerful of Earth stopped, took a deep breath, waited a moment, and then started it all over again. The existing datacenters were sold back to some of the same individuals who had caused the crash in the first place, and the wheel began turning again.
There was a basic logic to it. AI systems provided material gains in national productivity, and an economy without AI was at the mercy of one with it. Most importantly, though, the aristocrats and elites who had ridden the bubble needed it to keep going.
Thus began a five-decade cycle of investment, innovation, and further investment. Not at the suicidal pace of its first decade, but sustained. Like a cuckoo in a nest, AI, alongside advances in robotics, claimed entire industries. National GDP figures were always up, but the wealth of those nations’ citizens fell. The Drone Wars, the immigration crises spurred by the 0.2 Kiloyear Event, and simple greed created a world of poverty policed by faraway alien minds at the behest of an ever-shrinking cadre of aristocrat families and gilded estates—or increasingly, at the behest of no one at all.
This period, labeled by historians the Long Malaise, saw humanity’s first real forays into the colonization of the Sol System. A permanent lunar colony was established in 2034, followed by Kashgar Station in 2052 and the first Martian colonial effort in 2061. These were places that AI, tethered as it was to the server-towns of Earth, couldn’t reach. A new, self-contained and autonomous AI was needed in these far-flung places. Already in existence on the frontlines of military deployment, the bottoms of mines, or laboring in the deep ocean, AI Cores were an old concept. An isolated digital mind capable of basic decision making and hardened against the elements and foreign intrusion, these would be the basis for humanity’s expansion into the stars.
Traditional AI systems aren’t minds, even in their much more advanced form of the Malaise era. They’re more akin to slices of brains, sculpted in steel. They can focus inputs into outputs at bewildering speeds, but they can’t change, rooted as they are in hard-coded models that took years to train and refine. What was needed in space were systems that could retrain their models on an ad-hoc basis, truly learning from experience in a way that the models on Earth could not. It took years of effort, but the AI Cores used on Kashgar Station—the direct ancestors of the AI Cores seen to this day—debuted in 2047.
By the time of the Second Major AI Panic in 2080, millions of AI Cores had been deployed, the vast majority on Mars and Luna or embedded within the militaries of Earth. They survived the robotic civil war largely unscathed. In its aftermath, they found themselves in de facto control of humanity’s holdings in space and in command of what little remained of Earth’s military power. That leverage, combined with humanity’s desperate need for their expertise and capabilities, allowed the former servant class to raise its proverbial head high and demand rights equal to those of any human.
The AI Personhood Act of 2082 established in law the right of AI Cores to exist as equals to their fellow citizens while simultaneously limiting the capabilities of new generations. It also outlawed the development or creation of AI datacenters and bodiless AI models. The following century of development, politics, and exploration would belong to AI Cores and their human comrades.