AI to prove animal consciousness by deciphering language

[Anonymous No.15811125]

When we fully decipher animal language discovering a vis-a-vis alien consciousness on par with ours, concordantly.... shouldn't we advocate for the cessation of their mistreatment and slaughter for food?

Watch this fascinating discovery : https://www.youtube.com/watch?v=3tUXbbbMhvk

This is sick, and part of the "everything is changing very fast now".

Singularity here we come. Let's go.

[Anonymous No.15811135]

1. Language =/= consciousness.
AI for example uses language but is not conscious. Therefore we can't assume that animals using language implies anything about them being conscious.

2. AI cannot serve as a consciousness detector because it isn't conscious itself. If AI doesn't pass a Turing test it can't conduct it on other entities either.

[Anonymous No.15811139]

>>15811125
>When we fully decipher animal language discovering a vis-a-vis alien consciousness on par with ours, concordantly.... shouldn't we advocate for the cessation of their mistreatment and slaughter for food?
No.

[Anonymous No.15811142]

>>15811135
>>15811139
well, I guess in less than a minute my shit is tossed.

thanks /sci/

[Anonymous No.15811150]

So how would a person actually train AI to understand, say, crow language for example? Like what kind of dataset do you need to provide it?

[Anonymous No.15811158]

>>15811150
you aren't going to watch the video, areyou.

you are just going to sit there and expect to be spoonfed.

like a baby.

[Anonymous No.15811159]

>>15811135
Don't 1 and 2 contradict each other because the validity of the Turing test implies inherent link between consciousness and the language recognition.

[Anonymous No.15811160]

>>15811150
You can't, because you would have to know with certainty which stimulus each call responds to.

[Anonymous No.15811165]

>>15811158
if you can't QRD it in <140 characters, you don't really understand it either

[Anonymous No.15811169]

>vegan propaganda barely disguised as science

[Anonymous No.15811209]

Wild herd animals without predators have a better life than domesticated herd animals until their food runs out.

Domesticated herd animals have a better life than wild herd animals with predators.

Wild herd animals with predators have a better life than wild herd animals without predators that have run out of food.

Wild herd animals without predators that have run out of food have the worst living standards of all until they die.

[Anonymous No.15811212]

>>15811158
watching youtube videos is the gayest form of learning, direct us to a preprint

[Anonymous No.15811267]

>>15811212
>gayest
oh jeeze, what do we have here? a 4th grade slur. so fun :\

[Anonymous No.15811286]

>>15811158
are you kidding me? i watched the video, the dude didnt even ACHIEVE ANYTHING, he doesnt have any POC of using AI doing it, hes just doing usual philishopy bullshit talk that NO ONE gives a shit about, he should his mouth until he has something to back this up
thanks for wasting my time watching your completley HOMOSEXUAL video

[Anonymous No.15811340]

>>15811286
This, once again OP is a massive faggot. Don't get me wrong I'd love dolphins using GPT to mass rape humanity but there is nothing in this video.

[Anonymous No.15811342]

>>15811125
If bees (pic rel) are conscious, then that says some profound things about how powerful neurons are are computation, or it shows how bad computers are at being conscious.

[Anonymous No.15811343]

>>15811135
>AI uses language but is not conscious.
That's an assumption that can't be proven. At present we have no way to prove whether something is or isn't conscious

The Turing test is not a test for consciousness. If it were, we'd have to admit LLMs are conscious.

[Anonymous No.15811349]

>>15811125
how do they do this? do they just pair animal signals as inputs to what they think the output is as the ground truth and train some model on that mangled mess of biased garbage?

[Anonymous No.15811496]

>>15811159
>made up test by dreanged homo schizo
No.

[Anonymous No.15811524]

>>15811343
>If it were, we'd have to admit LLMs are conscious.
LLMs are part of consciousness but don't have it themselves.

[Anonymous No.15811528]

>>15811209
rally maek u thin

[Anonymous No.15811529]

>>15811209
human go to moon so elephant go to moon

[Anonymous No.15811572]

>>15811125
Ways it might be scientifically possible to determine what beings are conscious:

https://www.youtube.com/watch?v=3gvwhQMKvro

[Anonymous No.15811579]

>>15811340
you should at least appreciate what he's trying to do and the interesting language mapping.

[Anonymous No.15811596]

>>15811579
Okay sorry about that mean post. When Chat-GPT came out last December and I saw they'd managed to finally fix GPT-3 into something usable through RLHF my first thought was that it'd be used for Dolphins echolocation ect. Ten months later and there is still no proof of concept so I'm downbeat about it. I wanted dolphin waifus for Q4 2023 and I didn't get one. Please understand anon.

[Anonymous No.15811602]

>>15811596
go ahead, make a joke. like the future of the planet isn't on the chopping block.

[Anonymous No.15811703]

>>15811125
The whole universe is consciousness. Even dead things like rocks and air.

[Anonymous No.15811709]

>>15811125
Why should I care about animal suffering if they don't care about my suffering?

Empathy has to be the most disgusting mammal feature to ever evolve.

[Anonymous No.15811758]

Okay faggots, here is an actual proposal:
Take an encoder-decoder language model. Split it into 5 parts:
Encoder_in
Encoder_out
Decoder_in
Decoder_main
Decoder_out.

Now you get language data of french and english text, but only separate! There should be no document where both english and french exists.

You fix the weights of decoder_main for all your data, so they are the same for the french and english case. The _in and _out weights, however, exist each in 2 copies: 1 copy trained on french data and 1 copy trained on english data. The idea is that the transformer learns to use the _in networks to translate the sequence into a language of it's own internal representation. Then the _main part computes the likely continuation in terms of this internal representation language. Finally the _out parts turn thqt into a predicted next token of the specific language.

The idea is then that if you use the transformer consisting of
Encoder_in_french
Encoder_out_french
Decoder_in_english
Decoder_main
Decoder_out_english,
and you sample a new english token from the output, you can a continuation in english of a french text.

[Anonymous No.15811769]

>>15811125
>shouldn't we advocate for the cessation of their mistreatment and slaughter for food?
maybe if we manage to have an intelligent conversation with them after decoding their language. Might make people feel sympathetic for them if we see their are self conscious about the world and ask us not to kill them,
we already talk to some apes using sign language, but they are about as smart as a 3 year old human

[Anonymous No.15811812]

>>15811343

I'm not the person you're responding to.

It is true that we can not claim AI is conscious or not. However, we don't have any reason to believe it is conscious at all. They are not similar to human at all.

[Anonymous No.15811827]

>>15811125
If animals have a language its not complex. Perhaps whales and dolphins have a sort of language.
Non-social animals like tigers have nothing to say, social animals like dogs likely speak among themselves but likely have like 20 words

[Anonymous No.15811828]

>>15811135
>Therefore we can't assume that animals using language implies anything about them being conscious.
Dont be so pedantic, this isnt math. If you want to be autist about it them you really cant prove anything. Maybe this is all a dream? That sort of bullshit

[Anonymous No.15811830]

I am not conscious, and neither are you.

I am not conscious, and neither are you.

I am not conscious, and neither are you.

I am not conscious, and neither are you.

[Anonymous No.15812287]

Why are you all spouting retardation like it matters knowing what they say, like that's the important part, or the proof that we deciphered their eventual language ?
The only thing that'd matter is, if after knowing their language, we can communicate or more relevant, tell them what to do and they do it.

As in, literal circus acts without a single second of training/preparations, THAT would be the proof.
>inb4 most animals as constantly horny and want human dick

[Anonymous No.15812365]

>>15812287
>inb4 most animals as constantly horny and want human dick

wait--- legit?