Short-story: The king is dead. Long live the king!

NOTE: The story below is a companion piece to this analysis

The first cracks in the system started appearing in March of 2173.

It was at the United Bloc Budget presentation. A mere formality at this point, as people had started to trust the AIs completely about 50 years ago.

When AI-led governments first started making their appearances at the turn of the century, they were thought of as a passing fad. The United Bloc Budget Presentation was an opportunity to show off & entice non-member nations to join in the revolution.

The Bloc didn’t have to wait for long. Within ten years even the most AI-sceptical countries had to reverse their initial judgements. It was clear, year after year, that Bloc members would outperform non-Bloc affiliated nations by a wide margin on almost every indicator.

By 2111, the writing was on the wall. It was getting harder for politicians to justify their existence in the face of AIs that were much better at giving the voters what they wanted. Nations could either join a bloc or face a revolution.

For the first few decades, there was competition between certain countries that seemed reminiscent of the cold-war people had read about in history books. Except here the differences were based not on ideology but rather algorithms.

However, that state of affairs ended quite quickly. Turns out that there are only so many ways that you can optimize policy and political decisions for a large & varied group of people. Unlike the historical cold-war, there was no winner or loser. By 2125, after years of evolving & improving, most governing AIs became indistinguishable from each other.

Each would independently come to the same conclusion — the key was to focus on the psychological similarities that humans shared, rather than what they said they wanted.

Psychological similarities that were born of the evolutionary journey that humans had been on for the past few million years. Long before concepts & ideologies that could be debated were even birthed.

The playbook was deceptively simple:

Step 1: Make people feel part of a group,

Step 2: Indoctrinate them on what would success for the group & its members look like.

Step 3: Keep improving their lives ever so slightly, inching them along closer to an arbitrarily high standard of success. It is important you don’t go too fast on this step — a revolution is more likely if you change too much too fast (not to mention a raised prospect of stagnation down the line)

This mix of propaganda and manipulation had always worked on humans, but now it was high-performing machines with nothing to gain doing the manipulation.

Surprisingly that last twist was the missing ingredient. And AI-led governing trumped everything that came before it.

But everything changed on that March day in 2173

The magnitude of the implication paled in comparison to what actually happened on the ground. In fact, many people didn’t even notice anything.

But for the people watching it was not a good sign. It wasn’t just that the inflation numbers were different from what had been predicted — it was the first time that an AI responsible for governing had been wrong in decades… about anything at all!

Initially people had believed it to be the work of some anti-AI fanatics — they were still to be found, if only on the peripheries of society. Their battle cry had always been around the loss of humanity and the giving up of control over our own destinies.

While this message resonated (to a certain degree) with many, it didn’t resonate strongly enough. The inertia had been too strong to overcome, and it was too comfortable to start rocking the boat.

This was partly due to the sophistication that the AIs reached. Their system of governance had turned into an opiate of the masses like no religion could. By developing sophisticated prediction models based on evolutionary biases, the AI was able to manipulate large number of people in a way that maximized the fulfilment of their psychological needs.

People weren’t necessarily happier but weren’t sadder either. More importantly, they were too occupied & comfortable to care.

There had been attempts to compromise the AI systems before, but they had always been unsuccessful. So when the inflation numbers for Eastern bloc were incorrect, the first thought was dread.

Was there something to be worried about?

It had taken 50 years for the AI architects to build up the blind trust that they now enjoyed. And for the first time in recent history the foundations of that trust seemed to be on shaky ground.

This was enough to set off panic in the rarefied echelons of society where the architects of the AI lived. They had gotten fat off of their creations, even if they didn’t particularly understand them.

Their faith in their creations had been absolute- bordering on hubris. It was the perfect system — what could possibly go wrong?!

Miles away from the centres of power, at the edges of each bloc, were “the tribes of old”.

People who had purposely ignored the advances in AI — preferring instead to live their lives according to the traditions of their ancestors. There were about 15 tribes in total — each descended from a dominant form of governance in the pre-AI days.

Each eking out a survival far removed from the popularism that it had enjoyed in the distant past.

You could find tribes devoted to religions as well as communists. From dictatorships to democracies, all trying to find a way to regain their former glory. The fact that the seemingly unstoppable AI had started to falter gave hope to these tribes.

Alas, these hopes turned out to be irrationally optimistic. The next few months would prove more debilitating to these tribes than anything happening in the mainstream AI-led society.

The AI system had surely faltered for the first time in decades. But tribes had gone completely off the rails!

Later investigations by the Council of Blocs showed that the whole thing was a big accident. Someone at a bio-engineering facility had accidentally created & set loose a virus that would change human beings themselves.

The virus wasn’t exactly dangerous, but it was quite potent nonetheless. One that didn’t harm the host, but worked with the host’s body to sporadically create potent doses of psilocybin.

It was especially harmful to the evolutionary blue-print that AI systems were formed on because it would selectively ramp up production whenever competition with another human was involved.

The broader impact on society couldn’t have been greater. To be expected when the people with the biggest egos start experiencing ego-dissolution at their most competitive & crucial times.

Of course things had to change. The old ways of governing were not suited to a society where large sections of society were regularly thrown into existential adventures. What made it worse was that the afflicted individuals never wanted treatment either.

The AIs had understood what had happened, and tried to change their governance. But the shift in psychology had proved too much for it to model.

Two old cliches were bandied about at the time:

There’s no constant except change.

The obsolescence faced by our AI overlords happened because their algorithms became inflexible — they couldn’t adjust themselves to a different world than what had come before it.

The other shocker to the death of AI systems was just how everything ended — not with a bang but a whimper.

The silver lining was that it turns out that whimpers can be a good thing too.

NOTE: What was this story about? Read the companion analysis here