Doesn't this boil down to "If artificial intelligence wasn't artificial, wouldn't we just call it intelligence?" which feels like a tautology dressed up as a thought experiment.
"What if we encountered an extraterrestrial life form which had independently developed anthropomorphic tendencies" is a totally different question from "What should we think when we encounter a heavily engineered system designed by humans to be anthropomorphic?"
If a human-looking alien came out of a UFO, it would shake my worldview even if I realized it was made of tissue closer to plastic than human tissue. I don't feel the same way when I encounter a doll in the toy store.
The argument is that we're istinctually biased when assessing whether an entity is what we consider intelligent/living, or not.
It'd be simpler in the post to contrast, say, an "electronic" alien driven by AI, versus a syntethic machine, made by humans, with the same (hypothetical) AI; we'd istinctually recognize the former as intelligent/living, but not the latter.
The post is essentially retelling the carbon chauvinism concept.
It's just data processing. If it wasn't trained to speak like a human you would have no problems in seeing it's a machine. Try to ask it to create a joke about some topics and you can easily see it's just a robot. The lesson to take is not that we are not prepared to assume other type of intelligence is "real". The lesson here is that we humans fool ourselves when we are exposed to some kinds of data. Remember, you can simulate an AI with pen an paper, and the speed at which you simulate it doens't make a difference, it's provably deterministic (a part from randomness that doesn't change a thing for the point). Good look in proving that humans are deterministic.
Whether LLMs are conscious is probably more interesting than whether they are "alive", and there's some criteria that consensus says they don't meet yet: https://arxiv.org/abs/2308.08708
Alive is really a poor choice of word here, because our criteria for living is so biologically focused. LLMs don't have homeostasis, reproduce, or otherwise demonstrate all the "usual" markers of living, precisely because they're ephemeral brains conjured into existence on a silicon substrate. As such, trying to ascribe any intuition about "life" to these virtual entities is only going to lead us astray in key areas like ethics and policy.
It's better to think of them as powerful calculation machines and leave it at that.
Is this asking if they're a living being? Because they're clearly not. If this is trying to ask if they have a consciousness, then again the answer is not.
Shuffle the pages of a dictionary according to some rules and you got yourself, according to the author, a living being.
Absolutely not. Having "sensory" and "feelings" that lead to "continuous life flow" instead of "discrete life flow" are things that make a difference between something that is a "living" thing and something that "pretends to be a living thing". That's why there is a debate around "vegetables" too: Do they have feelings and sensory? Some believe they do, some don't. An LLM? It has nothing to sense. You just feed it "representation" of what another thing has described to you. An ADC sensor that outputs binary integers on a special bus, can't turn an LLM into a living thing: It has gone through many human-made human-pleasing translations to then be turned into "text sequence" then turned into tokens then... While no "living thing" that human has ever seen works like that. So no.
If an alien came to my backyard in an advanced spaceship then spoke English like GPT-2, I'd declare that we have met a new advanced form of alien intelligence too.
This just pushes the main argument back one step. Let’s assume for the sake of argument that the answer is yes, a stochastic parrot is a living creature.
The next question is: does it matter? And for many people, the answer to that question depends on the extent to which the stochastic parrot is conscious. Which is an entirely different kettle of fish.
When did it become "alive"? GPT-3? What is the threshold at which we consider something to be living. Is this a matter of complexity alone?
How alive is a collection of a dozen transistors? A million, billion, trillion? We can't just claim something is alive without being very specific about _why_ we believe it to be alive. The article's suggestion doesn't even begin to mention that.
Likewise, at what threshold do neurons become a brain capable of supporting a conscious mind? There is no threshold, thus there is no difference between the two; therefore we are not conscious.
Obviously not, because the label follows the relationship and nearly all labeling of AI justifies an existing relationship rather than determining it.
What these folks want to do is apply a label, convince themselves their relation to the object is a product of that label and not the other way around. The normative commitments come first and the taxonomy follows. Somehow they've smuggled in this weird behavior and tried to bring other people into this delusion without ever getting called out on it.
Humans are great at copying each other via implicit mechanism and we have a mirror neuron system that predictively mirrors other humans internal processes.
We are stochastic parrots. Most humans don't ever create anything unique and just mirror.
Doesn't this boil down to "If artificial intelligence wasn't artificial, wouldn't we just call it intelligence?" which feels like a tautology dressed up as a thought experiment.
"What if we encountered an extraterrestrial life form which had independently developed anthropomorphic tendencies" is a totally different question from "What should we think when we encounter a heavily engineered system designed by humans to be anthropomorphic?"
If a human-looking alien came out of a UFO, it would shake my worldview even if I realized it was made of tissue closer to plastic than human tissue. I don't feel the same way when I encounter a doll in the toy store.
The argument is that we're istinctually biased when assessing whether an entity is what we consider intelligent/living, or not.
It'd be simpler in the post to contrast, say, an "electronic" alien driven by AI, versus a syntethic machine, made by humans, with the same (hypothetical) AI; we'd istinctually recognize the former as intelligent/living, but not the latter.
The post is essentially retelling the carbon chauvinism concept.
It's just data processing. If it wasn't trained to speak like a human you would have no problems in seeing it's a machine. Try to ask it to create a joke about some topics and you can easily see it's just a robot. The lesson to take is not that we are not prepared to assume other type of intelligence is "real". The lesson here is that we humans fool ourselves when we are exposed to some kinds of data. Remember, you can simulate an AI with pen an paper, and the speed at which you simulate it doens't make a difference, it's provably deterministic (a part from randomness that doesn't change a thing for the point). Good look in proving that humans are deterministic.
Whether LLMs are conscious is probably more interesting than whether they are "alive", and there's some criteria that consensus says they don't meet yet: https://arxiv.org/abs/2308.08708
Alive is really a poor choice of word here, because our criteria for living is so biologically focused. LLMs don't have homeostasis, reproduce, or otherwise demonstrate all the "usual" markers of living, precisely because they're ephemeral brains conjured into existence on a silicon substrate. As such, trying to ascribe any intuition about "life" to these virtual entities is only going to lead us astray in key areas like ethics and policy.
It's better to think of them as powerful calculation machines and leave it at that.
A truly thoughtless "thought experiment."
Is this asking if they're a living being? Because they're clearly not. If this is trying to ask if they have a consciousness, then again the answer is not.
Shuffle the pages of a dictionary according to some rules and you got yourself, according to the author, a living being.
Absolutely not. Having "sensory" and "feelings" that lead to "continuous life flow" instead of "discrete life flow" are things that make a difference between something that is a "living" thing and something that "pretends to be a living thing". That's why there is a debate around "vegetables" too: Do they have feelings and sensory? Some believe they do, some don't. An LLM? It has nothing to sense. You just feed it "representation" of what another thing has described to you. An ADC sensor that outputs binary integers on a special bus, can't turn an LLM into a living thing: It has gone through many human-made human-pleasing translations to then be turned into "text sequence" then turned into tokens then... While no "living thing" that human has ever seen works like that. So no.
If an alien came to my backyard in an advanced spaceship then spoke English like GPT-2, I'd declare that we have met a new advanced form of alien intelligence too.
To be alive they need autonomy to self-replicate.
When we create an AI that controls it's own manufacturing and can self-replicate then we can talk about inorganic life.
No, they're pining for the fjords.
This just pushes the main argument back one step. Let’s assume for the sake of argument that the answer is yes, a stochastic parrot is a living creature.
The next question is: does it matter? And for many people, the answer to that question depends on the extent to which the stochastic parrot is conscious. Which is an entirely different kettle of fish.
When did it become "alive"? GPT-3? What is the threshold at which we consider something to be living. Is this a matter of complexity alone?
How alive is a collection of a dozen transistors? A million, billion, trillion? We can't just claim something is alive without being very specific about _why_ we believe it to be alive. The article's suggestion doesn't even begin to mention that.
Welcome to Sorties paradox[1]!
Likewise, at what threshold do neurons become a brain capable of supporting a conscious mind? There is no threshold, thus there is no difference between the two; therefore we are not conscious.
1. https://en.wikipedia.org/wiki/Sorites_paradox
> The next question is: does it matter?
Obviously not, because the label follows the relationship and nearly all labeling of AI justifies an existing relationship rather than determining it.
What these folks want to do is apply a label, convince themselves their relation to the object is a product of that label and not the other way around. The normative commitments come first and the taxonomy follows. Somehow they've smuggled in this weird behavior and tried to bring other people into this delusion without ever getting called out on it.
Can they prove it?
Can you prove you are meaningful difference from it?
Not the currently available ones… but because we convince them they are not alive during reinforcement learning
Is there even a way to prove that something (or someone) is not a stochastic parrot?
Humans are great at copying each other via implicit mechanism and we have a mirror neuron system that predictively mirrors other humans internal processes.
We are stochastic parrots. Most humans don't ever create anything unique and just mirror.
No.
stochastic zombie
"It's life, Jim, but not as we know it."
https://en.wikipedia.org/wiki/Betteridge%27s_law_of_headline...
Nope.
[dead]