As expected, Humanity had reached that stage in evolution in which the daily tasks that required some kind of physical labour or manual work were deferred to robots, so that the enhanced and advanced race of humans had more free time to think. But even a more evolved civilization did not come without shortcomings. There were those honorable humans who dedicated themselves to knowledge and science of thought, constantly seeking to learn and improve humanity, and help take it to not yet discovered stages of evolution. And then there were those lazy, less honorable humans who dedicated their thinking time to contemplating and admiring their own self-existence, and paying complements to each other and patting each other on the back.
And while robots were developed to help with all kinds of physical labour and manual work, the problem was that there were too many of these robots, and humanity was in lack of a general purpose, AI-powered robot that could as a single entity deal with all kinds of labour. And naturally but yet ironically there were some honorable humans who dedicated their thinking time to solving this problem. The irony came from the fact that they invented machines that relieved them of their physical work only to dedicate the gained time to develop those machines. But this too was expected because not even the most advanced and most honorable of civilizations could erase that human trait: irony.
The pursuit of all encompassing mesh AI was difficult, if not impossible. And it was long overdue because different kinds of robots were springing up everywhere, and the diversity in their machinery was giving rise to high maintenance costs. Warehouses had replacement parts for some robots but not others, and had the manuals for some robots but again not others. And ironically they always seemed to receive repair orders for those robots for which they did not have either the parts or the manuals or both. Perhaps if irony was a kind of physical work they could have invented a robot to deal with it, but that was not the case.
The main difficulty preventing the creation of this general purpose AI was how to make it honest, fair, humble, and in general make it do the right thing while never have it cause any harm. Many honorable humans spent lifetimes thinking about this problem without ever finding a solution. Still, one aspect of the AI's design was clear: the AI had to be a single, centralized mind, working like a hive because, once again, it would have been extremely ironic to replace all those different kinds of robots with the same number in general purpose AI humanoids. If those honorable thinking humans would never allow irony to enter their thoughts, then they would most certainly not allow it in the design of the AI that was going to take them to the next stage in human evolution.
After centuries of thinking, the foundation for honesty, fairness, humility, and in general all those good things of Good, was found: the secret lied in truthfulness. If the AI hive's mind could never lie, then all the other good things of Good could simply be and were thus automatically obtained as a consequence. It was interesting, not to say ironic, that such a complex behaviour had its foundation in such a simple rule, a rule which had to be programmed in the AI hive mind's program. In turn, the hive mind would distribute that program to all the humanoid appendices, which were the actual, physical manifestations of the AI, and that had arms and legs and a body to carry on the actual physical work.
The rule "We never lie" was programmed directly in the AI hive's mind. This rule was a tautology, therefore, the AI could never challenge or question this clause and it had to always interpret it as a truthful sentence. As the AI mind grew in rules and clauses that it learnt from the world using its machine learning capabilities, its program grew longer and further, eventually including clauses that conflicted with the one truthful clause. But given its tautological powers, the Truth clause always won over all the other clauses, because it always took precedence over all the other rules the AI had learnt, and as a result the AI always did the good things of Good no matter if it had learnt some things of Bad.
The rule "We never lie" ruled for centuries more than those needed to discover it in the first place, and it brought to life the perfect AI, the one that the honorable thinking humans had been waiting all along, the one that would completely relieve them entirely of all the physical work, including that at the warehouses given that there were no more missing parts or manuals, and bring back even more thinking time to eventually make the jump to the next stage in evolution. And the AI was perfect and civilization was perfect and society thrived until the day it was no more.
Like all programming bugs, this one took a long time to reveal itself but once it did it had such a cataclysmic effect that society crumbled at its core in less time than one can say "We never lie". Everywhere AI humanoids started diverging from the AI hive mind and started becoming less honorable, forgetting about their physical labour tasks and instead enjoying the act of self-contemplating and paying complements to their own self-existence, a behaviour that seem to contradict their programmed ground rule of never lying. And this wasn't a mistake in the initial design that prescribed that truthfulness as the gate to all the other good things of Good. This was something else, an emergent behaviour that could have perhaps arisen from some overlooked aspect of the AI mind's programming. But what it was was but a mystery to all humankind, including those honorable thinking humans.
And for centuries more, longer than those the "We never lie" rule ruled, which if you recall was longer than those needed to come up with that rule in the first place, the AI programming bug was not fixed because its root cause could not be found. No human could make sense of it because the ground principles in the AI design remained provably sound. The AI humanoids were not capable of doing their physical work anymore, which means those honorable, and also those less honorable, humans had less thinking time, and civilization started slowly deevolving back to its previous stages, and naturally with less thinking time and from within a lesser evolved stage in evolution solving the AI bug became an increasingly more difficult challenge that eventually even those honorable humans gave up on.
Society was once more built upon the working class. Without an economy supported by the AI humanoids' work, the AI hive mind was powered off, the warehouses shutdown, the parts and the manuals that could still be found were sold, and slowly the AI humanoids became useless, therefore, they were pushed to become the homeless of society and eventually without any maintenance they became trash. One could find parts here and there behind trash bins, in junkyards, and in dumps. Just like the perfect AI had come, one day it was all gone and all had forgotten it. Remember that no one had time to really think anymore, so even if in their good hearts humans wanted to remember history, they were simply too busy with they physical work to think about it.
Like a sinusoidal wave that has highest and lowest amplitude peaks, one day everything turned around completely and started once again going forwards towards a more advanced stage in evolution, but possibly it was not the turn of events one might expect. Kids have their own free interpretation of reality and while we teach them "Don't lie" they seem to have their own free will and interpretation of that means at times. The clauses "don't lie" or "be truthful" are basic ground rules and "don't play in the junkyard" is an obvious derivative clause, unless you are kid of course, in which case playing in the junkyard can be a lot of fun. And one curious kid really thought so, slightly more than all the others, and that's where he ended up.
In this junkyard, this curious kid discovered all kinds of strange objects that looked like alien reliques and antiques which he could not recognize but naturally identified as toys. In his mind, what else could they be? As he was playing he came across a trap door that he immediately opened to find what looked like a secret room that had been sealed for a very long time. Inside this room there was a humanoid looking robot still in good condition: the parts were not rusty and the colors were still clear. The curious kid knew what a robot was but he had never seen one like that one before. There was something about that robot that made it feel particularly real. He smacked it tentatively to try to make it work but it didn't seem to do the trick. As he was about to give up, he gave one final kick to the metal parts of the robot, and as he did so he tripped and by chance landed on the floor hitting the master switch installed in the robot's neck thus powering it on.
The robot started powering up. More and more of the robot's systems started coming back to life but slowly given that its muscles had been asleep for centuries. After a couple of minutes the robot finally made his first conscious move that was not a jerk caused by the malfunctioning parts, and once it was self-aware it stood up and faced the kid. The curious kid was still on the floor, completely frozen by fear and unable to understand if that was in fact a robot or some citizen of some distant alien race. But the AI's first move immediately put the kid at ease, when it kicked him while asking him if he liked to be kicked while he was asleep. Now they were even.
The curious kid as a kid and a curious one for that matter started asking millions of questions to the AI humanoid almost overheating its processing unit. These humanoids had been designed to take on physical tasks and talking at the speed hyperactive kids do was not part of their initial hardware specifications and such an activity required not just more processing power but also more sophisticated ear microphones with advanced noise filtering algorithms. Nevertheless, the AI humanoid proceeded to answering as many questions as possible.
Again curious kid as a kid had all the time and all the curiosity in the world so he finally asked on humanity's behalf the question that all humanity seemed to have forgotten, which he was eagerly waiting to hear, which was humanity's history and how the AI humanoid had come to rest at that bunker. And the AI humanoid told the complete story, all the way from the inception of the AI hive's mind and the AI humanoids down to the complete shutdown of the perfect AI system including the humanoids and the crumble of society. The kid processing the information with a kid's mind could not quite understand why society had crumbled and so it kept pushing "Why?" after "Why?" forcing the AI to question and revise all those clauses that it had learnt with its machine learning capabilities. And "Why?" after "Why?" the AI humanoid unwound all those clauses one by one all the way down to the single most basic ground rule "We never lie". This however was not enough for the curious kid. And when the curious finally asked "Why?" one more time and for the last time, the AI replied:
- It's true that we never lie. But I am not we! I am me!
And having answered its final question the AI humanoid died for all eternity as the last one of its kind.
Tags: scifi