If characters in the Sims become self-aware, what can they discover?

Great Debates
1

Basically, with more and more computer games adding AI to their characters, let’s assume that sometime in the future, some characters in a game of the Sims (or similar game) become self-aware.

And also let’s assume that they start getting curious as to where they are and why they are there. And let’s say they start approaching this question from philosophical and scientific points of view (with whatever “research” they are able to conduct).

What can they discover about their world?

I guess they can discover the basics, such as “if you drop something it falls” and “when there are clouds, it usually rains”.

But can they discover more fundamental stuff like what everything in that world is made of? Where their world is? Why their world was created?

To us, these questions have easy answers:

  • What is everyting made of: bits
  • Where is their world: inside Johnny’s computer
  • Why their world was created: because Johnny wanted to have some fun and Scooby Doo won’t be on TV for another half an hour.

But to people in that world, is it possible to find out these facts? And is it even possible for them to fathom these answers?

2

AI is probably the worst portion of actual consumer games (at the very best they are glorified lookup tables with random elements tosse in), so I don’t see Sim becomng self-aware anytime soon. But the problem for any such creature is that it’s unlikely they would see their existence in anything like the way we see it on the screen. However, they wouldn’t be able to see it as bits or computer language either, anymore than we can see our own neurons at work.

3

This is an interesting philosophical question. I’ve not played the game myself, although I think I pretty much understand the basic premise.

What would the Sims be able to discover about themselves?

Hmmm… well, for a start, are we accepting that they would sense the world around them in the same sort of way we sense ours? AFAIK, they don’t at the moment - Even the most sophisticated computer game AI doesn’t simulate senses properly - some do it better than others, but few do it by simulating photons bouncing around and being intercepted by a retina (OK, some of the better ones use ‘cones of sight’ as an approximation) - so what we see rendered on the screen might be a lot more detail than is available to the AI part of the program, should it become self-aware.
We might see a wall that the environmental software renders (for us) as being painted green, or decorated with floral wallpaper, but the AI might only have access to the information that it is a solid object through which it cannot pass.

Leaving that aside and assuming for a moment that the AI could perceive its environment as fully as we can, that would still leave it in a universe where objects only perform a limited range of actions forseen by the programmer - in the real world, glass breaks because of the properties of matter; in the Sims world, glass breaks because the programmer has instructed it to simulate breaking apart under certain conditions - if he hasn’t anticipated that, say, it should melt when subjected to high temperatures (and assuming there is some valid way to generate high temps in the Sims world), then simulated glass is at best going to do nothing when heated and at worst is going to malfunction in some way.

In our world, material properties work from the bottom up - they derive from the physical properties of matter - in the Sims world, they work from the top down - they derive from the programmer’s desire to make macroscopic objects behave in certain ways.

So a self-aware Sim is going to find his scientific observations rather frustrated.

4

An important question in any AI context: what is this thing going to be aware of, exactly? If we surmise that a CCTV camera is “self aware”, all it is possibly going to be ‘aware’ of is those specific images on the video tape. What are these Sims going to sense? Their virtual “world” provides so little sensory information that, to them, an ant is a veritable cognitive colussus, with its sense of sight, smell, direction, gravity and who knows what else, all tuned and adapted by millions of years of evolution.

As Apos suggests: all they’re going to be aware of is a modifiable lookup table.

5

Let’s assume yes.

Well, there is nothing to prevent a programmer to construct a world where microscopic “atoms” have certain properties and from those properties arise the properties of the macroscopic elements of this world (such as a glass)

As I said above, we can make a microscopically-based world.

Also, one thing to ponder about our world is, where do the “physical properties of matter” arise from? If you say, “well, that’s just the way they are”, one can say about a macroscopically-based simulation: “well, that’s just the way the properties of glass are”.

I think we can frustrate him/her even further by making the rules of the Sims such that, on a microscopic scale, things can be in two places at once, something can go from point A to point B without passing through the space in between, and our self-aware Sim can change the result of an experiment just by observing it :slight_smile:

6

We can assume that the simulation provides them with information about the “temparature”, how much light there is, who is around them, etc. Just the basic info that you and I need to get about our world.

In some sense, all we are aware of is what is on our retina, what comes through our ear drums, etc. We have certain sensory inputs and those are our interfaces to the world.
The self-aware Sim will also have a certain set of sensory inputs that describe its environment in great detail.

7

Another way to put it is, if we can have Sentient Meat , why can’t we have Sentient C++ objects?

8

Well, that’s just it: I’d suggest that humans, or ants for that matter, receive vastly more input than a simple temperature or incident light reading, and the “who is around them” is not supplied by the environment but actually requires computation which would make the operators of Deep Blue blanch in horror at its sheer complexity. The ‘basic info’ that you and I use to get about is anything but basic.

And what, precisely, is happening when we look at a 3-D ‘Magic Eye’ picture, or even that “vase vs. two faces” illusion? A human is as different to a camera as a real hedgehog is to Sonic.

9

Of course I agree with you in principle - I’m just suggesting that the “self awareness” that billions of years of evolution has equipped an ant (let alone us) with is decades, if not centuries, away.

10

Already explored here:
http://kidradd.com/

11

Indeed no, although you’re moving the goalposts - that is very far removed from the implementation of the Sims’ world. It would also require enormous amounts of processing power (orders of magnitude beyond what we currently have) to simulate an inhabitable world down to the subatomic level (and that’s ignoring quantum stuff).

In theory - yes, in practice - not yet.

No, that was my point; the properties of real-world objects are inherited from their materials; there’s a consistent, real-world explanation (in terms of electromagnetic fields, chemical valencies, crystal structures, molecular bond strengths etc.) as to why real-world materials behave in the way they do, whereas simulated objects can be programmed to behave in any manner you choose, even if it defies causal relationships; you could have a drinking glass that changes colour when it is picked up, or one that transforms into a fluffy kitten when you pour simulated orange juice into it and transforms into a planet-sized killer robot if you drop it.

There’s no way to formulate a meaningful understanding of a simulated reality when the properties are based entirely on the whim of the programmer; how would you define any useful physical laws? “Energy cannot be created or destroyed, except sometimes - for no apparent reason - it can”, “An object in motion tends to remain in motion, except sometimes when - for no apparent reason - it doesn’t”.

12

i think it’s safe to say that there is no possible way for a sim to perceive the outside world without outside intervention.

13

The way I see it, the first Sim that starts asking unanswerable questions dies horribly.

In public.
Inexplicably.
Spectacularly.

That ought to quell any introspection for a while.

Yea, verily, for I am wrathful, and vengeful, and I liketh not questions.

14

No disagreement here.

Maybe I was unclear in my OP. It was not my intention to limit the discussion to the actual SW game called “The Sims” circa 2005, but rather to a future version of something like the Sims that is vastly more complex than today’s Sims.

But the basic question remains: in this future, vastly complex simulation, if a sentient simulation object starts pondering its world, how much can it discover?

Can we formulate a theory of the knowable and the uknowable for these creatures?

Of course, the theory has to take into account the possibility that the world could be designed with specific mechanisms to thwart attempts by these entities to find out what is going on.

Maybe one can prove that unless you put specific mechanisms in place, these entities will always figure it all out eventually.

15

But why don’t you consider the properties of electromagnetic fields, molecules, etc to be based on the “whim of a programmer”?

In fact, I’m no physicist, but according to the latest theories, (Quantum Field Theory, IIRC), particles go in and out of existence all the time, something that is totally contrary to our everyday experience, and seems a bit arbitrary.

So, if you’re a programmer and you make laws that govern macro objects (e.g. all objects that fall turn into fluffy kittens) the self-aware beings in that world will just take those (bizarre) laws for granted, just as we take the (bizarre) laws that govern micro objects in our world for granted.

16

Why you silly goose, what did you think that spontanious combustion was? Besides, that wasn’t in public.

No, just kidding, I’m an atheist, there is no go::THUM::

Scott_plaid has been hit by a lightning bolt

Hah! You call yourself a god? That wasn’t any worse then the time I said Sha::KRAKTUME::

But really, the issue has been explored in fiction, most recently with video game heroes in a semi dramatic comic strip called “Kid Radd”. (no relation to the 80s’ cartoon) in which a 8-bit hero becomes self-aware, and again in a comical fashion involving the Sims in “The StrangerHood”. It is filmed in front of a live Sim audience. http://strangerhood.com/home.php

17

Why not ask yourself that question, SimBoy? :slight_smile:

18

A friend and I were once in the midst of an LSD experience when we decided to play Castlevania: Symphony of the Night. (Best sidescroller ever.) After tinkering a bit, he asked me the question, “If Alucard became self-aware, what would happen?” We approached it in two ways.

One was, “What would the character be doing?” We started on a long path of how we would become aware that he was aware. Suddenly he would not respond to controls as well, or he would respond sometimes but even when we weren’t using the controls he would move about without this input. We pictured little thought bubbles coming up onscreen and seeing questions and speculations about his surroundings. It was a sort of strange thing to think about in general, nevermind in such a state.

The other thing that puzzled us was, “How would we react to a sentient video game character?” Here we revisited the “how would we know” aspect. But we also focused on whether or not we would feel any responsibility towards Alucard. Would we feel compelled to leave the system on? Would we try and find a way to communicate with it? Horror of horrors, would we fail to realize that the improper control response was a risen consciousness and instead think something was wrong with the disc/system/controller and shut the whole thing off?

Ugh. Of course, from there the whole universe became a game and we were the characters and we were pleading with The Player to recognize our sentience and do what it could to make life better for everyone, to feel responsibility. But what if we failed to respond properly? Would The Player shut the system off…? My friend decided to sit there and do absolutely nothing in protest, while I pleaded with him to act normal lest The Player think his game is broke and power off the system.

Then I think we got distracted by the faux wood panelling. :smiley: Ahh, acid…

Still, I don’t mean to trivialize the discussion with the mention of drugs. I think it is really interesting and certainly the drug helped us approach the problem without that pesky everyday-worldview getting in the way.

But the OP was more about what the character(s) could discover. I think the answer is: not much more than the player could about that world. It might develop a better feel for object collision than we could, but maybe not. One problem the game character would face is a lack of interaction with like-minded individuals, unless we’re considering Sims[sup]2[/sup] to have all their characters become aware at about the same time. Even at that, there would still be underlying urges that would more or less compell them to behave normally. “Jobs” would not mean the same thing to them as it did to us, since they wouldn’t actually be doing anything for their money like we do. It would be interesting to ask what “going to work” was like for them, if they could even describe it. I think very little of our speech would be appropriate in any game. In fact, we might not even be able to translate anything, especially if they are limited to the small symbol set they are given in-game. What kind of communication would they have to set up in order to explore their world more in-depth? Would they develop a behavioral language of gesturing, would they manipulate furniture into different positions…?

Man, I’m going to be thinking about this for a while. Again.

Cool link Scott_plaid.

19

Your welcome. By the way, I forgot to put a link in the last post, so for convenience’s
sake. (Yes, that’s right I want the sake that belong to the personafacation of convenience. No, it’s not a spelling error.)

Ahem! Sorry, I got a little lost. Here are the Links again. For those who don’t want to scroll up.

Kid Radd

The Strangerhood

20

I know that several people have explored the issue of sentient beings living in simulations (e.g. Star Trek, The Matrix, …) and I am somewhat familiar with the “Simulation Argument” i.e. that we might be living in a simulation, and how should we live if we are living if we are living in a simulation.

But, I think my question is slightly different. I’m not interested in how a self-aware simulation object chooses to live. I’m curious as to whether there are any bounds to what the being can discover.

For example, this thread, Liberal seems to have used logic and reasoning to come to the conclusion that our world is not real, and that there must be a God and that only he is real (maybe this was in one of the links within that thread, and in any case my apologies if I am misstating what **Liberal ** is saying).

So, if we assume that the above is correct, then someone living *within *our system used reasoning to discover truths that describe our system.

Can a self-aware simulation object arrive at similar “truths”, and how much can they deduce?