Destructionator Domain

What follows is the final paper I wrote for my intro to philosophy class. It is pretty long (way longer than the teacher had in mind, I'm sure), but has some somewhat interesting stuff, so for anyone who is interested, here it is unedited and unabridged.

-------

Today I am going to explore a situation I find rather fascinating, which involves the ethics of creating throw-away copies of people. First, allow me to explain the situation. Then, I will get into the main portion of the essay, then finish it off by touching upon some tangential, but still interesting topics.

Consider a setting in the future, where people can be uploaded to computers and run as a piece of software. The specifics on how this is done is not particularly relevant to this discussion, but, let us assume it is a destructive process to the original mind. Thus, all we have left is the copy in the computer system.

This sounds like science fiction, since right now, it is. But it is an interesting question to the real world as well, since such technology will, barring some catastrophic event, almost certainly actually exist someday.

Moving on, once the person has been uploaded to the computer, then this person making a copy of himself should be a trivial operation - it is just a memory copy and adding the new copy to the running program list.

What I envision is similar to the Unix fork(2) system call. What that call does is make an exact copy of your program as it currently is in memory (as it runs) and adds it to the operating system's running list. The only difference between the two copies is one number: the first copy is returned the ID number of the second copy, and the second gets a zero returned instead. Using that single number, the programs may now branch to go about two separate, yet predetermined tasks.

It would also be interesting to explore a copy of a consciousness copied without this identifying return value. In this case, both copies would be identical in every way, thus would go about doing the same task in exactly the same way unless they can communicate the distribution of work through some other means, which may still get confusing - in a real computer program, the two processes might try to open a shared file to talk (which would be tricky to actually implement, since both would be trying to read or write to it at the same time). They might just generate a random number and proceed based on what it is (which, unless the random number comes from some external source, still won't work, since the two copies, being identical in every way, would have the same number generated!). Or, what would probably be the best way, would be the two copies would receive commands from a separate master program.

However, this seems to be a massive digression, since in all cases, the end result would be the same: one copy goes about one task, the other goes about another task (however they happen distribute the work doesn't really matter...) and then, when completed, one or both copies would be terminated. Not necessarily by some external entity either - programs normally terminate themselves when they have accomplished their specific task.

That is where things get interesting, since we aren't talking about normal computer programs - we are talking about people who happen to now live as programs. While it is perfectly normal for 'stupid' programs to kill themselves when finished with their work, obviously, no one would actually expect real people to do the same!

So, if we were talking about regular real people, this would be an open and shut essay. The reason this is interesting at all is that we are working with identical copies - a copy of someone, made by himself, fully realizing what he intends to do. That is, an intelligent being copies himself with the full knowledge that his copy will do a task then kill itself. The copy, being exactly the same, also realizes this in equal detail. Thus, the copy would be basically performing a consensual suicide.

Let's get on with it.
--------

I. Statement.

Already in setting the stage here, I've mentioned several things that could make tangential debates in themselves, and many of them would likely be very interesting. But the specific thing I want to explore is this: is it or is it not ethical for the original uploaded person to make copies of himself that will die? I'm kinda leaning towards 'yes'. Let's get started with the actual exploration.

II. Objections.

A.
First off, let's look at the 'no' side of things. There is a fair argument to be made here: namely, you are creating a sentient, living being (by a suitable definition of alive, of course, but let's go with cogito ergo sum here, which while not adequate as a general purpose definition, is certainly good enough for this specific discussion), with the sole reason of making it do work for you then summarily discarding it - in normal circumstances, this could sound like the worst kind of slavery.

B.
Second, consider enforcing that decision: what if the copy, being an intelligent being, decides it does not want to self-terminate when it completes its job? What do you do then? Do you forcibly kill it (that is, murder it)? That would certainly be unethical - you would be forced to continue letting it exist; keep the copy running indefinitely. This may be undesirable - in which case, you would have to never make the copy at all.

C.
Finally, what if the copy hates doing the work? Would you force him to do something he doesn't want to do? How will you compensate him, even if his life would be short? This is a person you are talking about.

I'll return to answer these objections in a moment.

III. The Defense.

Like I said above, there are so many tangential debates that could be had here, bit in this section, I will focus solely on the main statement above.

There are actually only a few cases where I would consider this to actually be an ethical course of action. In order for it to be a good decision, the person copying himself must fully understand what it means and be fully willing to do the work he would have his copy do and be fully willing to self-terminate after completing it. If the original is willing to do all this, the copy would be too, and thus no coercion would be needed. In any situation where he is not sure of himself, making the copy may result in enslaving or murdering the copy, and if there is a fair possibility of that, making the copy would be, at the very least, reckless.

So, making a copy like this would be shaky ground, but not something that should be prohibited. If the person doing the copy is prepared for it, you have a willing worker and a willing suicider (not a word, I know), brought into the world, and taken out of it, very cheaply and in an entirely painless fashion: the death of the copy would be instant; it simply would stop processing, thus, could not process any pain either. No suffering, all willingly, no outside interference nor dependence. That adds up to an ethical act.

IV. Back to the Objections.

A. (Recall that it was about creating a disposable slave.)
Normally, I see suicide (and other forms of self-damaging activities) as being unethical (another tangential debate), but in this case my normal arguments don't really apply because making the new person (the copy) is basically a free operation.

The reason I bring up suicide is the copy will be killing itself when it has completed its work. As an intelligent being, all other things being equal, it would have the right to control its own life. (Again, all things are usually *not* equal - normally, suicides affect other people too: friends, family, etc., whose lives you have no right to harm. Here, a short lived copy would have no such worries.) Thus, the copy would be free to terminate itself ethically.

Moreover, given it is a copy of the original being made, it would have the same likes and dislikes as the original. Assuming the copy is being made to more quickly do some work that the original would have done anyway, it should willingly do the work. It isn't going to be enslaved nor murdered, so it is OK to create it.

B. (Recall that it asked 'What if it doesn't want to die?')
This objection is based in great part on the assumption that the copy has free will. It would not, since free will does not exist (tangential debate again!). If the original would be would be willing to do the job and kill itself when completed, the copy would also be willing to do so, guaranteed. (Though, if I am wrong about free will, doing this would potentially be one of the hard experiments that could help attack my position.)

So, since there is a 0% probability of this situation occurring, it does not alter the ethics of creating the copy, given the caveat that the one copying himself truly understands and accepts what the copy means.

C. (Recall it asked about compensation and the copy's willingness to work)
This ties into what I talked about in the defense: if the original knows what he is getting into and is willing to do it (meaning he copies just to help save time, since two 'heads' can be faster than one), then the copy will also be willing to do it. He would need no compensation; the copy knows he exists to do this job, and successfully doing it should be enough for him. If it isn't, the original should indeed rethink his decision to spawn a copy of himself, but it is not enough to warrant a blanket statement of judging the practice unethical.

--------

Tangential debates:

Note to professor: I mentioned above in the main part of the paper that there are several little things I could talk about at some length; things that are interesting debates in themselves. While not strictly fitting into the paper's format and scope, I nevertheless felt they were too interesting (and fun to write) to simply leave with at best a short parenthetical comment above, and have chosen to expand upon some of them below. Naturally, since they do not really fit into the format, I don't expect them to be graded (one way or the other), but am handing them in simply because they are somewhat fascinating and you might enjoy reading it when you have some spare time.

1) On the existence of free will

This is one that seems to strike a nerve with a lot of people with whom I've talked about it, but their responses often seem to be primarily logical fallacies of the appeal to consequence variety - 'if what you are saying is true, X would happen, and X is undesirable, therefore what you are saying must not be true'. But, of course, that doesn't follow, so I don't accept it as a rebuttal of worth.

First, let's try to define it. Free will is when someone or something can make its own decision, based on his specific situation, that cannot be deterministically predicted ahead of time.

My argument is pretty simple: there is no (known) mechanism that allows free will, and it is not necessary in explaining any observations. Therefore, we assume it doesn't exist, until shown otherwise. (In science, we are always open to the possibility that we are wrong and some new discovery will change things.) Some will argue that quantum fluctuations could account for it, but the brain works (as far as we know; much more research on this subject is certainly justified) on primarily the chemical level - too big a scale for quantum effects to make any large, regular differences.

Hell, quantum stuff is as best we can tell truly random - it wouldn't so much account for free will due to my clause in the definition saying a free decision must be made with the full knowledge of the specific situation anyway, even if the hypothesis otherwise had no flaws. This isn't necessarily a killer though; the random effect might tip the deterministic decision making process one way or another, thus making it non-deterministic. But the scale and frequency of them seems to remain the real deal-breaker here.

Moving on, I've heard people claim that AI is impossible because no current computer can have free will, and a 'true' AI would need free will, thus, no true AI could run on any hardware. The fatal flaw here is of course it assumes two things: it assumes free will exists, and it assumes we, and anything like us (alive by definition two - see below) must also have it. Sometimes, the same people who put forward the random hypothesis discussed above will refuse to apply that in the case of computers: we can build hardware random number generators that operate on principles similar to that (note that while I don't think it will work on the chemical scale of the brain, electrons and transistors in computers are a whole different beast. We know it works there, as devices have actually been built with (apparently) true randomness.). But the anti-AI camp will say that doesn't count as free will! Applying a different standard for people and machines seems to be indicative of some bias in their definitions and arguments.

I've actually argued this point with Ph.D. students in Computer Science before on the Internet. They have pages upon pages of mathematical proofs to support this along with several published references by other CS Ph.D's. All that is nice, but seeing how it is all founded on an unproven assumption, all that supporting math; all their arguments, are simply irrelevant.

Free will isn't a matter of philosophy nor mathematics. It is a matter of pure science - it doesn't exist until shown otherwise with real world observations and a real world mechanism that accurately describes the observations. Naturally, this does require a good definition, but the one I proposed above should work pretty well, and seems to be generally agreed upon by people to whom I've spoken.

The consequences of accepting this fact I feel are entirely irrelevant. Using them to attack the conclusion is of course a logical fallacy, but there might be some merit in discussing them from a philosophical viewpoint on their own. I'll quickly explain why I feel it is irrelevant, then move on, leaving the rest of this debate as an exercise to the reader :-)

It is irrelevant for this simple reason: does it change how we would live our lives? Does it /actually/ have an appreciable change? Not really. Some little things might want to be adjusted, but certainly nothing catastrophic.

The big thing often brought up is ethics. It is often argued that if people have no free will, then they cannot honestly be held accountable for their actions. I don't buy this. Two reasons: we can continue to pretend that free will does exist. After all, it certainly appears that way day to day on the people scale; a free will based model of reality, even if not actually true, works for most things. Second, even if you don't pretend, society must go on. People doing unethical things are bad for everyone (pretty much the definition of 'unethical') and should not be permitted to continue doing so, so you discourage it, or punish them, etc., for the benefit of a continuing functional society.

Of course, since we don't really have free will, we aren't actually making those decisions, but again, the free will model of reality is certainly useful.

2) On the morality of suicide (and other self-destructive activities)

I briefly summed up my position on this above: it is unethical because it affects other people.

The pro-self-destructive activities (a catch-all term that includes suicide) crowd argues that a person has a right to do whatever he wants with his own body. Even if we grant this, the fact is that self-destructive activities do more than affect just the person's self. It has effects that can be seen in many people.

Another, less personal effect on other people self-destructive activities have is related to that of debt. A regular person is very expensive to create: children cost their parents upwards of $500 a month, each, and even more on society as a whole (direct benefits from going to school, indirect benefits from getting medical, fire, police protection - all the things everyone gets and takes for granted from the government).

If someone voluntarily did something to make a group of other people miserable, this would fall under unethical in several systems, including my own (which is a slight modification on utilitarianism, where one of the important distinctions from what most people think when I say utilitarianism is I measure harm as bad as the primary concern rather than happiness as good, but you get the basic idea).

The pro-self-destructive activities people have at least two arguments here: a) the reactions of other people were not intended by the activity, and could not be controlled by the person in question and b) the person in question has no duty to those other people. I (obviously) don't buy into either of those. For argument a, I would say intent is irrelevant. The fact is that you did cause this pain, and you should have saw it coming (it doesn't take a genius to figure out your family is going to miss you when you die). The effects happened, and you knew it was not only possible but quite likely, so it is unethical to do the self-destructive activities.

The keen counter-debater may dream up the following situation to use against my logic: consider a fair lady with several suitors who all want to marry her. She will of course realize she can only pick one, at the exclusion of all the others. Is this fair lady being unethical by marrying just one of the suitors, leaving the others saddened by this choice? If not, why can the same logic not be applied to the self-destructive activity performing person? Of course, our gut says that it is not unethical for her to make a choice, and I would not argue with it in this case. But what about applying the logic? Does it just not work?

Maybe it does apply; let's break it down. It is very important to remember that ethics is not a black and white study, nor are its conclusions nor each step that goes into it. An ethical judgement is a complex decision made by weighing many complex parts. Before I get into the break down, let me concoct a similar scenario with a self-destructive activity.

Consider a kid smoking marijuana* despite his friend's protests. From the protests, he can reasonably conclude that continuing to use this drug will cause hardship in his friends, much like the fair lady rejecting most the suitors. However, the difference here is the marriage of the fair lady and her chosen suitor is in itself a positive act; it won't be harming either of them. On the other hand, tt has been proven that using drugs like this will directly harm the individual, but this one probably won't kill him outright; he can potentially make a full recovery from the damage it causes. Thus, it wouldn't as extreme an example as suicide, but extending the logic to the extreme example should be trivial.

* Let it be noted that this author** absolutely despises the use of recreational drugs with a passion that cannot be matched by the eternal wrath of the flames of a thousand suns. I will try to conceal any bias I might have here, of course.

** Let it be further noted that this author absolutely loves referring to himself as 'this author'. Okay, enough of that, time to get back on topic.

So what we have in the drug example is the user's friends having to deal with watching their friend hurt himself, and the user of course being hurt. In the suitor example, we have two people turning out quite well, and the others not gaining something for themselves, but not really losing anything either. However, they do get to see the one they were suiting turn out happily. The drug user is just bringing pain all around, the fair lady is bringing a little pain, sure, but mostly, happiness. Thus, the self-destructiveness falls under unethical and the fair lady falls under ethical under the same logic and the same process, defeating the counter argument.

Continuing on to argument b from above: it is certainly stronger, especially if you subscribe to duty based ethics (which I generally do not; there are some exceptions (ethics aren't black and white), but I normally phrase it in other terms). But, consider the following:

One could argue that society is making an investment in each child, and investors hope to get a return on their money. In this case, they would hope that the child grows up to become a functioning, contributing individual to the whole. You could also look at his consumption of public resources (such as schooling) as a kind of loan given to him from society with the hope and expectation that he will eventually pay it back in the form of his own taxes later in life, to help the next generation the way the previous helped him.

When someone ruins or takes his own life, he is telling voluntarily throwing away that investment made into him, telling the investors that their work into him was just a waste of effort. He is defaulting on that loan through his own actions. Almost seems like theft, but not quite. Anyway though, throwing away other people's hard work for your selfish acts falls under unethical with my system, and ignoring your duty to pay back this debt could be crammed into going against Kantian ethics as well (there could be a debate here with the motivations of the self-destructive activities and how they fit in or do not fit in with the categorical imperative, but I'll leave that as another exercise to the reader, since I am running short on time...).

3) On the definition of life

You have two definitions of life: one (the correct one) is the biological definition, which covers everything from bacteria to people. The other is the one I am using here, referring to life in a 'thinking' sense - like I said above, 'cogito ergo sum' pretty much sums it up.

I'm not really sure life is the word that should be used here, since it doesn't really fit the correct, biological, definition, but it is the best word I can think of right now that gets the point across.

Working out a more specific definition for this is actually pretty difficult. At what point would you call an artificial intelligence 'alive'? For human uploads, I think it is easy enough to just give it to them. After all, the exact same entity in a biological body was also considered alive (by both definitions), so it seems perfectly reasonable to grant him exactly the same status in an artificial body.

In the general case though, there is much room for discussion. My own definition used to be anything that can think on a level at or above where it can act on its own. Anything that fits this should be treated like people, since they would be ethically equivalent (that might be another tangential debate... with a specific enough ethical system, you could construct an artificial life form with greater than human intelligence, but none of the qualities that are ethically relevant, thus making it disposable. This is actually being worked on in the real world, and if I had the time, I'd write up my thoughts on it.) This definition has several problems - what, specifically, do I mean by acting on its own? How would you go about measuring (reliably recognizing) that? What about something that has comprehension in some other category, but not here? Heck, what is acting on its own, given that free will doesn't exist (see above)?

It just doesn't work without more modification.

A better definition might be one that is more broad: perhaps, any thing that is aware of its surroundings and can react accordingly to it. But this also fails: a simple electronic motion detector or automatically focusing camera could fall under it.

Maybe if we mixed in the definition of intelligence (basically the definition I am think of is 'the capability to, in a structured way, alter free variables in your environment to help you achieve your goal'). In fact, mixing in the proper definition of life too would maybe work: biological life basically covers things that try to maintain their internal state through a series of complex chemical reactions, reacts to its environment, and is capable of reproducing itself.

So, our new second definition of alive is: any entity that reacts to its environment, tries to maintain itself (generally stay alive), and is capable of copying itself. This still falls short; a self-replicating factory (NASA has some designs for them) would pretty much fit.

Let's remember what this definition was for: primarily, ethical debates. For reproducing stuff, the biological definition will probably always be the most proper. So, since ethics is a focus, we could add to the previous a clause like 'and is ethically relevant; that is, it is capable of the things your ethics system cares about'.

This would probably mean that is is self-aware, at least mildly intelligent, and capable of feeling pain or sadness.

I think we have something; let's stop with that for now and move on.

4) On the feasibility of uploading

Back in the free will discussion, I mentioned that I've argued with computer science students over the feasibility of artificial general intelligences. I've also spoken to people who think it is indeed possible, including one expert from the UK currently working on actually creating one. This expert thinks that doing human uploads are a bad idea, and makes several good arguments there. To sum it up, a computerized human could alter himself to almost a limitless extent, which would probably have unpredictable and very likely bad consequences. After all, look at how many humans try to alter their mind right now with things like drugs, and look at the bad effects that has.

However, he does think it is technically possible, and proposes two methods to accomplish it. One is slowly replacing brain cells with artificial equivalents, which when done, will now have a copy of the original brain that can be plugged into a computer. The second consists of using imaging technologies to scan a brain into a computer, likely destroying the brain in the process. The second is the method I assumed above, simply because it takes the original copy out of existence as it goes, simplifying some aspects of the above debate.

Both these seem technically feasible to many people. But, there are also many who think this would not possibly work. The arguments usually have one fatal flaw mixed in, often hidden beneath layers of extrapolation: there is a hidden assumption in there that people are more then their physical selves.

Like the free will discussion, this in itself is simply a matter of evidence, or rather, the lack thereof. There is no evidence whatsoever to show this additional thing exists; no observations that depend on its existence to make sense. And, of course, also like with free will, we are always open to new evidence that shows otherwise (which the upload experiment could provide), but until that evidence arises, we must assume that there is nothing beyond the physical world, and thus, from everything we know, making an exact copy of a brain would result in an exact copy of the individual's memories, personality, etc.

Yes, I am saying people have no souls here, so taking the classic arguments against the soul and applying them here should work for the most part against many objections that may be raised. To recall a few: where do these souls come from? When a fetus splits into twins, what is the soul situation there? Even if you assume it does exist, why could you not apply the answers to the previous two questions in the creation of artificial life or copies of life? With or without soul, the last part there seems to be a moral blow to the anti-AI argument in that respect.

Moving on, actually running this copy will require simulations of the brain's processes either in some form of hardware or in software. When the computer is not running, the brain image will lay dormant until it is run again. From the perspective of the uploadee, the time he is not running will be no time at all. (Now, turning him off, making a copy, discarding the old, and running the new copy on new hardware brings up a debate often seen in Star Trek fandom, relating to their fictional transporter, where many fans argue that every time the characters beam up, they are being murdered and recreated (based on some fan's interpretation on how the device works - zap something into atoms, beam it somewhere, put it back together in exactly its original form). But in both cases, I would not call it murder at all; it is merely discontinuity of consciousness. Yet another debate that can become quite lengthy, but I must leave it here as I am behind schedule.) All this should be technically possible though, since the sum of a consciousness is not any more than its constituent data.

As of right now, actually doing this procedure is science fiction, but with more research on the brain and continuing work in computer and software engineering, the time should eventually come when the current hurdles are jumped, and then, we'll see for sure. Oh what an interesting day that will be.

Adam D. Ruppe

Our host moved us to a different server, but I kept posting to the old server over the last week. If you could not see any new posts over the last few days, that is why.

I just did data migration, so we should be good now, and nothing should have been lost.

I have decided to set up a blog here on LibArc to be the current repository for my random nonsense not generally useful to create debate (as that would go on the forums), nor generally useful for reference (as that should go on the wiki).

I haven't had the real drive to maintain my own static pages for quick randomness, like I prefer, thus the blog shall do that task.

My categories will eventually be the following

:

• General - here, I will put anything that does not fit into the other categories.
• Software - this will be about computer software, including what I am currently writing and using. I will also talk about software development ideas (such as Open Source).
• ASE - this will be talking about my original science fiction universe

I will surely think up more as I go.