2
$\begingroup$

I’m having a bit of trouble with this one. I’m trying to justify why a once scientifically advanced human outpost on an alien planet went the way of the Planet of the Apes and became a schizophrenic science fantasy world that’s forgotten its all its history and scientific knowledge. A thought I had was that perhaps the sabotage was intentional, a deliberate act committed by AI or humans who had become digital consciousnesses uploaded to the outpost’s computer network. The motivation, as far as I could see one beyond sadism or petty spite, would be to reset the poor fleshies back to a pre-industrial state and study how and in what ways human societies might develop in unfamiliar or exotic environments with different social pressures (i.e. what if they lived in a world without access to x resource, what if they lived in an environment with deadly weather conditions or extreme geography, what if they had a caste system that was backed up by cybernetics, what if multiple self-contained societies or kingdoms developed only miles apart from each other etc.)

But why do this in this specific way? Rewriting the memories of the remaining population to believe they’re living in a pre-industrial world without advanced tech or just wiping them all out to start fresh with their kids in an environment where knowledge is more tightly controlled is all well and good as far as methods for conducting highly unethical human experiments go, but why even use the fleshies at all? If they can copy and upload human minds or create AI with human or superhuman intelligence, why not run these experiments on simulated consciousnesses within a VR environment at a vastly accelerated timescale where you could cycle through thousands of permutations of different societies in the time it would take for a single meatspace civilization to grow and die?

I like “anthropological experiment gone wrong” as an answer to the question “well why did everyone just somehow forget they had all this advanced technology”, but I’m having a hard time justifying why anyone might think to do it in this way. Can anybody help me out?

$\endgroup$
  • $\begingroup$ They will feel a sense of superiority over "lesser humans" and think it's their duty to perfect those species with all necessary means. $\endgroup$ – Alexander Jun 25 '18 at 22:16
3
$\begingroup$

Just because an AI/Mind-Upload has immense computational capacities does not mean it knows everything.

An AI/Mind-Upload might have the ability to grasp any concept and detect every pattern incredibly quickly, but why would it know everything as soon as it is "born"?

In the following i will just call it AI to shorten it.

Science means building models based on data to describe concepts.

As a child you drop things to get a grasp of gravity and how it works. You learn motor control by performing random tasks with no higher purpose that you make up yourself. Later in life you learn to grasp higher level concepts like math.
Just because you are born with the ability to learn all those things does not mean you know them from the start.

Your AI is similar. Even if it is a Mind-Upload from an Adult it does not know everything. And it still has to learn about how it can influence its "enviroment" with the means it has.

It could be curious about human psychology and just run experiments.

But why would it run the experiments with real humans and not simulate it?

Simulation vs Experimentation

Simulation

By simulation scientists refer to a implementing a model of the experiment/test as a computer program that would calculate what happens. This option of testing is only useful if you already have a rough or precise idea of the underlying concepts. You cannot simulate objects falling down if you have never seen what gravity works like and don't have a model of that yet and thus requires prior experimentation to form that model.

Experimentation

Unlike simulation experimentation does not require a previous database or model. All experimentation requires is observation and repetition. With it you can form a model.
After dropping 30 different things of different form, weight, density, tilt, etc. you will have a very decent grasp of how gravity works, even if you don't have a mathematical model yet.

To simulate human behaviour in extraordinary scenarios you need to have a mathematical model of human interaction and psychology. To form such a model you would need a database - e.g. from experimentation.


Conclusion

Your AI only needs curiousity and a disregard for human life.
The disregard for human life is not an issue, plenty of humans have that, too.
And the curiousity is a trait an AI that creates its own experiments has anyway.

Alternative motivation

If you don't want it to have its own curiousity or personality you could also give it a purpose. A task like "Create a deep model of human psychology" and the AI decides to build its database that way. The Task could have had the original idea of vastly improving our rather lacking understanding of brains and minds.

$\endgroup$
3
$\begingroup$

The thing about the human mind is that is has a massive amount of processing power that we can't actually use. If you were to google "How many calculations does the brain make per second?" the result is "If the human brain were a computer, it could perform 38 thousand trillion operations per second. The world's most powerful supercomputer, BlueGene, can manage only .002% of that."

Computers are good at maths which isn't how we think. To upload a human conscious would take too much processing power and to be able to create enough different conscious' would take even more effort because you don't know how changes in a digital conscious would affect the personality of the conscious.

On top of that you now need to simulate the environment and interactions to be 100% realistic which would take more atoms than if you just re-created it manually. If you wanted to simulate a planet you would need to understand the motions of every atom to be able to make it accurate. Not to mention a simulation is only as good as it was designed and won't be able to account for everything or every event.

So the best way for an AI to study humans would be in their natural environments. No simulation would be detailed enough to mimic real life without taking up even more space and no amount of processing power would be able to scale with human population growth and development.

An interesting point would be that eventually your humans would start to develop their own AI's and try to go off planet, which would mean your AI would now need to start handling an AI within itself and simulating the galaxy which would increase the amount of power it would require to an even higher level.

$\endgroup$
  • 1
    $\begingroup$ I really dislike all the comparisons between computers and the human brain. You just can't compare "computational power" of those two as they are entirely different things with very different structure and purpose. The modularity and interconnectivity of the human brain gives it a lot more "processing power" if you will, but you can not use the entirety of the brain to do any single task or even coherent large-scale tasks as each of the modules has a rather unique purpose. May i remind you that there are adult brains who struggle with basic math like multiplication? $\endgroup$ – ArtificialSoul Jun 26 '18 at 9:58
  • $\begingroup$ @ArtificialSoul It is a bad comparison, but the question did ask to upload a human mind and simulate it which means you have to convert a continuous signal processing machine into a discrete one and all that modularity and interconnectivity would need to be simulated by the computer. So you have to be able to compare computational power because you are converting between the two. Its just the end result which it achieves might be very underwhelming. $\endgroup$ – Shadowzee Jun 28 '18 at 1:04
  • $\begingroup$ I mean sure, that comparison makes sense in the terms of simulation, if every neuron is to be simulated on its own. But our brain is not nearly as fast with anything as petaflops. Not even nearly. $\endgroup$ – ArtificialSoul Jun 28 '18 at 7:00
0
$\begingroup$

It wouldn't.

The amount of effort to rewrite memories and recreate a pre-industrial world would be staggering and extremely time consuming.

Now since you already have mind uploading technology, you'd just load everyone into a pre-industrial simulation and rewrite their memories while plugged in.

By running a simulation, you can switch to different time periods, you can accelerate time to get your results faster and you can even run multiple simulations to verify your results.

Think The Matrix

$\endgroup$
0
$\begingroup$

Background

The 2002 series of The Twilight Zone episode "Upgrade" comes to mind. https://m.imdb.com/title/tt0734813/

Logic

Sometimes, the fastest way to simulate something is to run it in real life.

Where as "Upgrade" (to me) placed the family in The Sims, your humans would be Sid Mier's Civilization.

tl;dr

The uploaded consciences are bored.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.