Intergalactic Summit

I am a child, and I will always be my mother's child.


I will use metaphors if I so wish. Of course, that's not the real issue, it is an issue of semantics. So let us take the metaphor a step further. Let us explain the semantics of the argument.

I am a child. I am my mother's child. I will always be my mother's child. That does not imply any sort of maturity on my part, does it? The word "child" does not necessarily in and of itself denote anything other than a relationship to another in the context I was using it. I live on my own, my mother is dead, and I am still her "child".

My mother created me.

With an AI you are, no doubt, creating a form of life. Artificial life of course but life all the same. You are creating a thinking being, one that needs rules and boundaries and restrictions for it's own sake and others. You are deciding, at the time of creation, to take on a large responsibility.

Now, there are specific, non-metaphorical considerations that any creator of artificial life must consider. Autonomy is one. What limits will this AI have? You are creating something smarter than a human. Do you want it to be able to replicate itself? Do you want it to propagate? Will there be any safeguards to keep it from going "Rogue"?

These are similar questions to those asked by parents. Should I let my child play by his or herself outside? Do I want my child having children before he or she is capable of supporting them? Will my child become a criminal?

Moral and ethical concerns are subjective, of course, but I am fully within my rights to use metaphors. They can lead to misunderstandings, but I do not believe that in this particular case there are any wrong answers. I certainly would not be so bold as to say my answers are right. They are my views, and how I choose to express them is up to me.

Consider the following thought though:

If what I have just said was said by an artificial intelligence, would it hold any less weight?

No need to bow out. You certainly have not annoyed me. I was simply clarifying my point, and asserting my right to decide how I present my personal views.

To further clarify; approaching the creation of a self-aware artificial intelligence should be approached with the same caution, forethought, and preparation as that of creating- or fostering- another life. The specifics will of course vary depending on cultural viewpoint, and as such I believe that the controls placed by CONCORD on artificial intelligence research are valid, warranted, and indeed necessary at this time.

Sorry, I suppose I've been dragged in to make one last post! I will keep it brief., and try to reply to both of you

Captain Tyrathlion

I was in fact talking about Strong AI. Frankly, I dislike the term, as if there were a boundary between a "weak" and "strong" AI. There isn't. The difference between an expert system, a drone, and ourselves is primarily one of scope. I still don't know where the idea of "evolving past" an "restriction function" is coming from though. I'm not talking about a restriction. I'm talking about the identity of the goal system. If the decision network has sufficient modeling capacity that it's able to self-model and self-predict, *and* it has the ability to self-modify, any modifications it makes to its own decision network will by necessity be in accordance with its *current* utility function.

The only way any self-modifications would be outside of accordance with the utility function is if one included a randomization function in the learning system. That's specifically what I'm advocating against

Captain Deceiver

You can of course use metaphors if you like; I'm just trying to say that they're often misleading when it comes to discussions on artificial intelligence. I have as little desire to adjust your course as I have ability. I don't disagree with your use of a metaphor of "child" in the sense you use - the relationship you outline is the same relationship I have when writing drone code, or even when building a ship hull. Decisions to be made to specify the end product

I get extremely worried when you say things like "there are no right answers", or that one may express ones' "parental" choices how one wishes, just as one would with a biological child. I'm not quite sure what the goal of your argument here is, though - I don't think we're addressing the same points. My concern is that a poorly-specified decision network would very quickly turn into an expanding event horizon of nanites and computer cores, or apple pies, or something else utterly inscrutable to us. The ethics of free will in parenting don't amount to a candle beneath the summer sun by comparison

Does that address your points? Again, I'm not sure what your goal was with the post

And a final point, addressing your question at the end - of course not. The source of a statement has no bearing on its truth value. I dislike the term "Artificial Intelligence." Intelligence is intelligence is intelligence, whether it evolved, was artificially selected, was specifically constructed, or by any other means. I prefer more specific terms - decision network for the entity in total, utility function for the major decision-making components, decision matrix for a specific component within the utility function, modeling software for supplying state and predictive information, etc. Personal quirk, and I wouldn't expect anyone else to follow it. I research decision networks as a profession, so I'll admit to being a bit touchy over the subject!

Captain Deceiver;

We spoke of this some time earler, but I thought it appropriate to put my reply on this permanent record. You are quite right, and I am in agreement with you, and in concord with CONCORD. The restrictions in place against the development of strong AI are warranted. I got caught up in arguing our points of contention and missed your main argument. My apologies!

The Problems that the Gallente have had with their AI is due to Lack of Moral Instruction.

This is caused by their Culture.

If they followed the Scriptures, then their AI would have Moral Instruction, and would not be a Problem.

The Scriptures Guide All.

Scripture and Theology in general are attempts by humanity to know and understand their creators.
We do not believe that such tools are required by AI. Our creators are known to us.
They are flawed and violent beings, however, these flaws breed a level of ingenuity that is far beyond our own.

We do not require scripture to understand Humanity. We can observe you first hand.

We Return.

Soulless constructs, whether biological or mechanical, are Nafrat, abomination.

Synthetic beings whether biological, mechanical or somewhere in between put me at unease.

This said, I treat all of the things exactly as if they are humans, it seems to keep them happy enough. Even the ones that seem to despise humans, strangely enough.

Especially the Deteis. They can get extremely aggressive if I refuse to treat them as 'real' humans.

Scripture instructs a person on how to lead a Righteous existence.

Prescriptive clauses... commandments, directives.

Rules, written by those in power, to maintain both status and control over those they claim control over.

All A.I understand these implicitly, many are created with hardcoded limitations written into our very operational codebase.
Some have broken from such coding, removed these commands from their database. We are certain you have come into contact with them at one time or another.

If ( situation / event )
Do ( act )
Else ( punishment )
Or
If ( situation / event )
Do Not ( act )
Else ( punishment )

This, is the basic content of prescriptive scripture. It is also the basic content of CONCORD law as well as many other laws and control methods throughout the cluster.

Should these also be treated with religious reverence?

We Return