[ShinyLeftPad]

A hammer is not alive. You are the one who called LLM "alive" and used the word "enslavement".

The logic "it is our creation" does not mean we get to abuse and enslave it.

Even with animals that we don't consider sentient we generally agree that unnecessary cruelty is bad and we usually tolerate cruelty only if it is a threat or a source of food (basic necessity) which LLM is not.

If you believe it is 100% not alive and not sentient then that's fine, but your original comment said something different.

[jvanderbot]

You have to read one comment above mine to find a statement about how they may someday become more than simple tools. Let me quote it:

> The pope treats them exclusively as lifeless tools, and while I think that this is most likely true for current LLMs, we should think more about the limit of what is going to happen in the next 20 years than being overly focused on the current capability level. To me it seems rather likely that we will soon have machines that plausibly experience consciousness.

The insinuation is person-hood may emerge. My assertion is that even though some might believe (e.g., the top level discussion) that they are more like persons, this is wrong.

My statement is that, no matter what, they should _at most_ qualify as domesticated labor animals. But I'm not even willing to go that far if I can help it, and think anyone who goes further than viewing them as text generation programs has a mental disorder.

I should have simply added "at most" in the early part of my comment...

> The logic "it is our creation" does not mean we get to abuse and enslave it.

A hammer does not experience abuse, nor enslavement, so it's irrelevant for the purposes of discussing AI. The humanization occurred _before_ my comment, and I was pushing back on it, not introducing it.

Is that more clear? Does that chain the steps together better? It's great that you picked on the weak analogies in my comments, but I can only dance around core thesis so many times.

[ShinyLeftPad]

I react to your ethical position and logic.

Saying that it's OK to enslave and mistreat some living thing because we enslave and mistreat others (animals, in the past humans). For sure mistreatment of animals is a BIG issue and the main acceptable justification is human survival, because they're threats or food. Applying this logic to something that is not threat and not food is flat out wrong.

> My statement is that, no matter what, they should _at most_ qualify as domesticated labor animals

Outright cruelty to horses is taboo in many places and even horse racing is controversial.

Of course if it's a fully mechanical tool then there's no concept of slavery or cruelty. But we are morally obligated to do our best to find out if it's more than a tool. The article makes a point that we have no consensus how it works and whether it's truly just a tool. It definitely behaves like a human, and we all usually agree that human deserves rights to live and be free and such, without them we suffer and it dehumanizes whoever takes them away. But in any case, if there's a chance it's not just a tool, even if it's at animal level, we shouldn't just say "even if it's alive it's fine to enslave it because it's for sure no better than an animal and we have a history of animal abuse already"

[jvanderbot]

I reject your axiomatic assertion that AI is a living thing, and therefore qualifies for any comparison to animals (an assertion I was rejecting all along).

I think we're only talking because you think that I believe AI is a living thing and choose to advocate for mistreating it. Or, perhaps you are arguing because you think I want people to be cruel to animals. Also not true.

I'll try again. If you compare the conditions that animals are raised in, they are horrific, agreed, and I love many animals, and yet we as humans mostly accept animal's plight in the world at our hands.

I'm saying, I would accept those as the best conditions for an AI, just like I don't heat my toolshed to keep my rake warm in the winter, because it is my property and I care nothing for my rake's comfort, and my neighbor has no right to claim that I should care better for my rake (or car, or desk, or fridge).

> But we are morally obligated to do our best to find out if it's more than a tool.

Ah, another disagreement: No, we are not. You may feel that way, and do your best to discover the inner consciousness of anything you like, but we are under no obligation to entertain each object as though it were conscious unless some legal rights have been established for it. We may choose to because we are good people here and there (opt-in morals), but in the end I would drown, by hand, any animal in the world in front of its owner and parents to save a child. That hierarchy sets children above every other animal because of their basic humanity. I would probably destroy most tools to save an animal. And I would probably delete (or reset) an AI before I would destroy any material thing. That's the real hierarchy.

You and I are having an interesting discussion, but we are polluting this forum by going off topic and using very strong assertions with each other. I would be happy to continue this by email, which is available via my profile.

[ShinyLeftPad]

You are distorting my words a lot.

> I reject your axiomatic assertion that AI is a living thing

It was your original assertion but it's 100% irrelevant anyway. I argue against your logic regardless of whether it is alive or not.

> we are under no obligation to entertain each object as though it were conscious unless some legal rights have been established for it

We went through this during slavery. There were no rights, and then people realized there should be rights.

And "each object" is a straw man. We were talking about beings to which the concept of cruelty and slavery could apply, like an animal. The assumption was that it could be the case with LLMs too. Without that assumption we should not be having this argument.

[jvanderbot]

> We were talking about beings to which the concept of cruelty and slavery could apply, like an animal. The assumption was that it could be the case with LLMs too. Without that assumption we should not be having this argument.

Agree. Cheers to a mutual resolution.