Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Three Inverse Laws of AI

susam.net

Post reply on HN
Print view View original

61–70 of 388 posts

Re: Three Inverse Laws of AI

#61

Post by Eisenstein »

quectophoton wrote: post #30

> Humans must not anthropomorphise AI systems. Can someone explain why this is a bad thing, while at the same time it's a good thing to say stuff like "put a computer to sleep", "hibernate", "killing" processes, processes having "child" processes, "reaping", "what does the error say ?", "touch", etc? To me that's just language, and humans just using casual language.

The people who advocate for not anthropomorphizing are afraid of the implications of integrating these systems into society with implicit human framing. By attributing to AIs human qualities, we will develop empathy for them and we will start to create a role for them in society as a being deserving moral consideration.

Re: Three Inverse Laws of AI

#62

Post by devmor »

largbae wrote: post #41
miyoji wrote: post #6

I strongly disagree with this framing. It's patently insane to demand that humans alter their behavior to accommodate the foibles of mere machines, and it simply won't work in the majority of cases. Humans WILL anthropomorphize the AI, humans WILL blindly trust their outputs, and humans WILL defer responsibility to them. Asimov's laws of robotics are flawed too, of course. There is no finite set of rules that can con…

The article offers practical advice to go along with this framing, like configuring AI services to write/speak in a more robotic tone. I think that's a decent path to try.

This is actually one of the things that made LLMs more usable for me. The default tone and style of writing they tend to use is nauseatingly annoying and buries information in prose that sounds like a corporate presentation.

Re: Three Inverse Laws of AI

#63

Post by kokojambo »

Great article. Fully agree. Ai is not something that can hold responsibility, a human overseer is always required. These overseers are to be held accountable. Note however that these overseers are also highly prone to blame ai when mistakes occur in order to avoid judgement and punishment. When a person says "ai did this/that" always wonder who guided that ai and how and if proper supervision was given.

Re: Three Inverse Laws of AI

#64

Post by bjt »

wolttam wrote: post #25
cobbzilla wrote: post #22

Earlier quoted context omitted.

We have invented a new tool that can cause great harm. Do you see any value whatsoever in promulgating safety guidelines for humans to use the tool without hurting themselves or others? Do you not own any power tools?

Of course there is value in promulgating safety *guidelines*. But we cannot guarantee those guidelines to always be followed.

Guidelines on their own probably won't be taken too seriously.

But other things will:

- Liability rules

- Regulations that you get audited on (esp. for companies already heavily regulated, like banks, credit agencies, defense contractors, etc)

If you get the legal responsibility part right, then the education part flows from that naturally.

Re: Three Inverse Laws of AI

#66

Post by glenstein »

>Humans must not anthropomorphise AI systems.

Yes, but. Starting with my agreement, I've seen anthropomorphizing in the typical ways, (e.g. treating automated text production as real reports of personal internal feeling), but also in strange ways: e.g. "transistors are kind of like neurons" etc. And the latter is especially interesting because it's anthropomorphizing in the sense of treating vector databases and weights and so on as human-like infrastructure. Both leading to disasters that could be avoided if one tried not to anthropomorphize.

But. While "do not anthropomorphize" certainly feels like good advice, it comes with a new and unique possibility of mistake, namely wrongly treating certain generalized phenomena like they only belong to humans. Often this mistaken version of "don't anthropomorphize" wisdom leads to misunderstandings when it comes to animal behavior, treating things like fear, pain, kinship, or other emotional experiences like they are exclusively human and that thinking animals have them counts as "anthropomorphizing." In truth the cautionary principle reduces our empathy for the internal lives of animals.

So all that said, I think it's at least possible that some future version of AI could have an internal world like ours or infrastructure that's importantly similar to our biological infrastructure for supporting consciousness, and for genuine report of preference and intent. But(!!!) what will make those observations true will be all kinds of devilish details specific to those respective infrastructures.

Re: Three Inverse Laws of AI

#67

Post by miyoji »

cobbzilla wrote: post #22
miyoji wrote: post #6

I strongly disagree with this framing. It's patently insane to demand that humans alter their behavior to accommodate the foibles of mere machines, and it simply won't work in the majority of cases. Humans WILL anthropomorphize the AI, humans WILL blindly trust their outputs, and humans WILL defer responsibility to them. Asimov's laws of robotics are flawed too, of course. There is no finite set of rules that can con…

We have invented a new tool that can cause great harm. Do you see any value whatsoever in promulgating safety guidelines for humans to use the tool without hurting themselves or others? Do you not own any power tools?

I see value in promulgating safety guidelines for power tools, sure.

There's another comment comparing LLMs to shovels, and I think both that and the power tool comparison miss the mark quite a bit. LLMs are a social technology, and the social equivalent of getting your hand cut off doesn't hurt immediately in the way that cutting your actual hand off would. It's more like social media, or cigarettes, or gambling. You can be warned about the dangers, you can see the shells of wrecked human beings who regret using these technologies, but it doesn't work on our stupid monkey brains. Because the pain of the mistake is too loosely connected to the moment of error. We are bad at learning in situations where rewards are immediate and consequences are delayed, and warnings don't do much.

I guess what I'm really saying is that these safety guidelines are not nearly enough to keep us safe from the dangers of AI that they're meant to prevent.

Re: Three Inverse Laws of AI

#68

Post by jdw64 »

soks86 wrote: post #59
jdw64 wrote: post #12

I understand that AI output is generated from statistical and representational patterns learned from a vast amount of data. My understanding is that, during training, the model forms high-dimensional internal representations where words, sentences, concepts, and relationships are arranged in useful ways. A user’s input activates a particular semantic direction and context within that space, and the chatbot generates…

I still haven't read any of his work, but wasn't the point of the Three Laws of Robotics that they in fact _didn't_ work in the story presented in the book?

[dead]

Re: Three Inverse Laws of AI

#69

Post by beepbooptheory »

miyoji wrote: post #6

I strongly disagree with this framing. It's patently insane to demand that humans alter their behavior to accommodate the foibles of mere machines, and it simply won't work in the majority of cases. Humans WILL anthropomorphize the AI, humans WILL blindly trust their outputs, and humans WILL defer responsibility to them. Asimov's laws of robotics are flawed too, of course. There is no finite set of rules that can con…

Do you consider all things broadly called "ethical" to be similarly a waste of time? Even if we lived in a world where everyone always behaved unjustly, because of some like behavioristic/physical principle, don't you think we would still have an idea of justice as what we should do? Because an ethical frame is decidedly not an empirical one, right?

We don't just look around and take an average of what everyone is doing already and call that what is right, right? Whether you're deontological or utilitarian or virtue about it, there is still the idea that we can speak to what is "good" even if we can't see that good out there.

Maybe it is "insane" to expect meaning from something like this, but what is the alternative to you? OK maybe we can't be prescriptive--people don't listen, are always bad, are hopeless wet bags, etc--but still, that doesn't in itself rule out the possibility of the broad project that reflects on what is maybe right or wrong. Right?

Post reply on HN