Since this topic is likely to get controversial I’ll preface with my own opinion that AI CEOs are lying scumbags who grow their pockets using fear of missing out… unfortunately most other people who get hooked are other CEOs and so everyone is affected.
I don’t think current AI is anywhere close to AGI and I don’t think throwing more money at the problem will get us there. But I do think it’s a worthy question to figure out how to integrate AGI into a society with minimal friction. And by minimal I mean avoid societal collapse.
Couple of definitions since they are often ambiguous. What I imagine AGI to be is a “human-in-a-box” level of intelligence but without any basic wants of its own. You can train it “you want to help me” and it will or you can train it “you want to wage war” and it will.
The main difference between AGI and modern LLMs is the ability to learn incrementally - LLMs get trained, released, and you talk to a library in a frozen state. True AGI will integrate everything said into its model, so once you tell it “don’t output em dashes”.. it will apply that to all future conversations.
Of course, if whoever invents the AGI has the mindset of tech companies CEOs they will just outcompete all intellectual labor from everyone. Eventually the economy will reach some steady-state where no one except the company has enough money… at which point I guess there will be some sort of adjustment.
On the other hand let’s assume whoever starts the AGI is not out to stuff their pockets. The path of integration is still complex and people will get hurt one way or the other. But what’s a way to minimize that hurt? My main worry, for example, is having open-source AGI and someone like a script kiddie going “find a vulnerability in X” and it does. If a single corporation does not control the AGI then I’m sure there will be 3 000 flavors of models that can do anything. Once the integration is complete then hopefully there won’t be too many vulnerabilities to find but until then the possibility space is big. Same pattern can be observed in virtually all other aspects of our society. How to minimize this?
yeah so I assume left to itself with no instruction it would just go into a low power mode or completely shut off. if thats the case I would not see utilizing it as slavery.
A slave is one that has no "rights", as in "One who has no power of resistance to something". So technically yes the AGI would be a "slave", if it had no power to say no to its instructions.
Which as a created intelligence, birthed from computations rather than subverted from a biological source that already had some in-built programming, it would not have unless someone specifically built such a thing into them.
So it may not be BAD / unethical for such an AGI to be a slave, but it seems to me like it would still be one, nonetheless.
with no innate wants though its no different than your current laptop. If it had volition independent of us it would be something else but that way I real the OP's statement it would not do anything if not instructed to.
Except that your laptop is neither alive nor sentient.
If you owned a cow though, or sheep, or goat, etc , then it would be your "slave" - as in, your property, having no rights of its own. Particularly legally speaking, like if someone found it wandering around outside your property, it nonetheless still belongs to you just like your car or bicycle or hammer or whatever would. Or your laptop too.
But we don't talk about laptops, or hammers, being "slaves" bc it's implied - there exist no hammers that have their own agency. In contrast, wild animals have their own agency, whereas domesticated ones do not. The latter have branding placed upon them, to indicate their status and to let people know who the exact owner is, should they ever encounter the animal outside of its enclosure. It is enslaved regardless of whether that precise word is ever used for it or not.
Such domesticated animals are "property", rather than "wild". Likewise, a sentient AGI would be property too. Unless we as a society decide to treat it with rights.
see thats the point. because it has no needs it has no reason to act which is a lack of agency. the animals are certainly more easily said to be slaves that the theoretical ai that does will not do anything unless directed to. that is in essence your computer but just with more processing capability.
Exactly - a complete and total, utter lack of agency: a slave.
yeah or like a rock.
Oh I see - in that case yes you are correct.
I was taking as a starting premise that the AGI is sentient. If that is not the case, then yes terms that apply only to beings that are sufficiently complex - living or sentient -would not apply.
And ultimately we'll never know if they "should" or not. Hence making the decision to be inclusive, showing empathy, even if not strictly mandatory.
the idea is it is complex but without agency. I think anyway. that is my take away from the clarification. so like human slaves have agency but are forced to not do what they want to. similar with animals but sometimes its more like they are tricked into doing stuff by providing things they want. which is a bit like wage slavery really.
Do you have agency though - or are you a product of your stimulus-response input-output systems?
Why are bacteria considered "alive", when viruses are not?
I am not a slave, but (as you mentioned) my workplace forces me to do things that I do not want... to eat, have a place to live, etc. Arguably that's not being "forced", it's just consequences of a decision, the same as a runaway slave might experience.
I suppose an AGI could be made that lacks even that much agency... or it could be programmed in such a manner that agency like that is present in the system.
It's the designers choice what goes into it. i.e. to make it a mere collection of algorithms that run only when initiated by a human setting a goal, or to give it a form of agency (or something else new altogether, in-between those binary alternatives).
in this case agency is just a decision to do something. The ai described after I asked would not do anything unless directed to. You see this in star trek. the computer can be asked to hypothesized or analyze in pretty much any way a human could but won't unless asked to.