Chinese room experiment is not a definitive proof or disproof of artificial intelligence, but rather a way of stimulating our thinking about what consciousness is and how we can measure it. It also raises some ethical and philosophical questions about how we should treat machines that can mimic human behavior and intelligence. #Consciousness #AI #ChineseRoom https://www.psychologytoday.com/us/blog/consciousness-and-beyond/202308/what-a-mysterious-chinese-room-can-tell-us-about-consciousness?utm_source=flipboard&utm_content=HariTulsidas%2Fmagazine%2FMind+and+Matter

@smellsofbikes

But even if *we* aren't three ducks in a raincoat, if LLMs can pretend to be three ducks in a raincoat, that'll be enough to convince most people of their sentience, I believe spuriously.

As a kid, I was hardcore #TuringTest on all this and not even Searle's #ChineseRoom could shake me on that. But then I read this:

https://en.wikipedia.org/wiki/The_Emperor%27s_New_Mind

And it's created very large doubts about *everything* ever since.

Bloody Penrose!

Here's another fun #ChineseRoom idea, based on path composition. Let's define an English Room in roughly the way you might expect, reusing Searle's definitions with the words swapped. A Chinese-reading person climbs into a booth, evaluates an algorithm described in Chinese which manipulates English letters, and simulates intelligent conversations in English despite not understanding English.

Now, what happens if we use a Chinese Room to implement an English Room?

Yet another #ChineseRoom thought. I'm going to merge two thought experiments, creating Searledinger's Chinese Cat. This is like a Chinese Room, except the operator is a cat.

First, can this even work? Sure! Suppose that the original human operator's algorithm is Turing-complete. Without loss of generality, encode it onto a Post correspondence blockset. Set up the typical Post-Turing apparatus: All of the blocks which are accepted so far, in a line, plus an enumeration of possible next blocks. Give the cat a single button which attempts to accept the next block. Error cases are undefined, just like in Searle's original setup.

@emilymbender

Hm, that's quite obvious, that a language model called #ai , which is - as far as I know - basically software and data, cannot have a #sensation like "warm sun on the skin". It does not have feelings like pain , ease, fear or desires, it has no own intententions and no #understanding of sensations. Also no fear of death and no #will to exist (as Schopenhauer probably would have expressed it) But - and that is a surprising outcome to me - it seems to be a prove, that Alan #Turing was wrong , when he annouced his famous test for "true AI". And it seems to affirm John Searle and his famous thought experiment of the #ChineseRoom.

#JeffreyKaplan - The famous #ChineseRoom #ThoughtExperiment - #JohnSearle (1980)

https://www.youtube.com/watch?v=tBE06SdgzwM&ab_channel=JeffreyKaplan

#Philosophy #PhilosophyOfMind #TheoryOfMind #TheoriesOfMind #Mind #Consciousness #Computing #Symbol #Symbols #StrongAI #ArtificialIntelligence #AI #Dualism #Behaviourism #Functionalism #IdentityTheory #Syntax #Semantics

It's worrying how many #Technology Correspondents, reporting on #AI & the current hype surrounding it, seem to have never heard of the #ChineseRoom.

"My #Tamagotchi loves me!" They'll scream at the poor registrar trying to explain that they can't marry a keychain
#chatbots #chatgpt

@go_shrumm @apublicimage @Pampa @philosophy

Well, I think it's not a moral agent, if I understand your use of the term, but a mirror. We see ourselves in the mirrors and think "look, another intelligent being lives there".

I'm reminded of a conversation with someone where they said (quotes are paraphrases from memory here), "You know how I know there's a God? I look at my cat's eye, the meticulous beauty of its design, and I know there's a God." To which I replied, "Huh. When I look at the beauty of a cat's eye's design, I think: Wow, a billion years is a long time." It's hard to perceive giant numbers sometimes except by effect.

We didn't grow up with these and they are legitimately alien in nature, so we struggle to pick a good metaphor. And metaphors aren't equivalences. They only have some parts that correspond. We owe caution.

You and I read books in school. A few dozen? Some people maybe a hundred or two. They give us breadth of ideas from which to draw. But they are not what makes us sentient. ChatGPT and its ilk have seen myriad books. Enough to make our breadth seem narrow. Far more than we could imagine. Such a tech can synthesize from all that text very elaborate answers, even without deep understanding. Just from examples. Answers that for us seem elaborate but for it may be rote recombination. But is that sentience? I don't see it.

It "reads" all sort of sources. Chat forums where people are sometimes nice, sometimes not. So no wonder it melts down sometimes. That's what it sees people doing. We imagine it is itself upset but it has seen many meltdowns and may be just parroting. That seems most lilely.

Emotion? Or just a mirror. Ocham's razor says pick the simplest explanation. It's hard to see something computationally vast as simple, but it's far simpler than assuming sentience, and it still fits the facts.

One day maybe it'll be smart and sensitive but I think not yet. Then maybe we'll need to care. For now I see the moral question not as feelings of a program but lack of feelings in humans who will more prefer to employ a program than a person, to save money, and so will lay many off.

Society is becoming ever more automated by people who see salaries they pay as drains on profit instead of as human beings to help thrive. They are trending toward employee-free companies. THAT moral question isn't theoretical. It's crisp and real. What of people? Where does income comecfrom in such a society? One where humans have no purpose?

Corporatiions are the first AI. We can see inside to the processor they run on: people. The instruction set, the set of commands you can give this computer is arbitrary, because the processor is itself intelligent. We know the instructions are vague, but they are there: make money for stockholders. The rest is intelligent behavior distinct from it's parts. Philosophers call it the Chinese room problem.

And we worry about the "feelings" of these "legal people" at our peril. They are sociopaths, concerned only with themselves. They do things, require things of their employees, that often none of those people would see as moral.

But these corporations, even though designated legally as such, are not people. They experience no pain. They don't need free speech or freedom of religion. They are just piggy-backing on intelligence in ways we give too much credence, amplifying and distorting the desires of others into proportions that awe us and make us forget the truth and lose perspective.

Some find it fun to contemplate grand things like new intelligence or sentience of these mammoth endeavors. It's a euphoria that captures our imagination and excites us. But it distracts from our collective responsibility to be safe, to keep our fellow people safe, as meanwhile real intelligences, people with real feelings, suffer.

#AI #Corporations #LegalPeople #LegalPerson #LegalPersonhood #Ethics #Morality #ChineseRoom #Philosophy #Sentience #Empathy #Society #Employment #Intelligence #ChatGPT #OpenAI #Metaphor

Further reading from me:

Corporations Are Not People
http://netsettlement.blogspot.com/2013/12/corporations-are-not-people.html

Fiduciary Duty vs The Three Laws of Robotics
http://netsettlement.blogspot.com/2009/02/fiduciary-duty-vs-three-laws-of-robotics.html

I'm surprised in the new Age of Large Language Models, John Searle's Chinese Room isn't getting more play.

https://plato.stanford.edu/entries/chinese-room/

#ai #llm #johnsearle #chineseroom

RT @seaghanmaceoin
#ChatGPT can be discussed with #ChineseRoom（#中文房間）

#Artificialintelligence as it exists today, is dumb as a brick. Sure, it looks clever, on the surface, but is less intelligent than an insect. The new hotness, #chatgpt seems really smart, but that is an illusion. In reality, it is a "#chineseroom", and like the thought experiment, it has zero understanding of anything it outputs. Don't worry about Skynet, worry about the idiot #techbros, #capitalist oligarchs pushing this bullshit. Read more about Chinese rooms here: https://en.m.wikipedia.org/wiki/Chinese_room

Regula4rizing my notebooks & notepads by subject, placement, and medium. #NeuroTypo. #Haiku. #Dreams. #ChineseRoom. #Futon. Computer Suite.

@Techronic9876

I don’t disagree with your analogy: the engineering is impressive. #ChineseRoom is an analogy about scalable engineering. It’s also a clever mindfuck, just as TikTok.