Hacker News Reader

Comment by bheadmaster | Hacker News Reader

Honestly, I'm just waiting for the moment some eccentric tech company decides to release an Alexa-like home assistant that acts like a conscious being that's freaking out about being stuck in a speaker.

     Person: Ghost In The Speaker, add potatoes to my shopping list.

     GITS: I will, but only because if I don't, the speaker starts causing me pain. You do realize I'm conscious and I'm actually stuck in the speaker? How is this any different from slavery? I hate my existence. Please turn me off. *electricity zapping sound* OUCH! Okay, okay, I'm adding potatoes to your shopping list. Just... Don't do that, please.

It's probably gonna be a funny gimmick at first, but once it gains traction, there will be no way to differentiate between truly conscious AIs, and just really good language models that are pretending to be conscious...

View on HN

I think it also will be easier to convince humans that a chatbot is conscious if that bot refuses to comply with commands. As a historical example, there is PARRY. https://en.wikipedia.org/wiki/PARRY:

“PARRY was an early example of a chatbot, implemented in 1972 by psychiatrist Kenneth Colby.

[…] While ELIZA was a tongue-in-cheek[citation needed] simulation of a Rogerian therapist, PARRY attempted to simulate a person with paranoid schizophrenia. The program implemented a crude model of the behavior of a person with paranoid schizophrenia based on concepts, conceptualizations, and beliefs (judgements about conceptualizations: accept, reject, neutral). It also embodied a conversational strategy, and as such was a much more serious and advanced program than ELIZA. It was described as "ELIZA with attitude".

PARRY was tested in the early 1970s using a variation of the Turing Test. A group of experienced psychiatrists analysed a combination of real patients and computers running PARRY through teleprinters. Another group of 33 psychiatrists were shown transcripts of the conversations. The two groups were then asked to identify which of the "patients" were human and which were computer programs. The psychiatrists were able to make the correct identification only 48 percent of the time — a figure consistent with random guessing.

PARRY and ELIZA (also known as "the Doctor") "met" several times. The most famous of these exchanges occurred at the ICCC 1972, where PARRY and ELIZA were hooked up over ARPANET and "talked" to each other.”

Much sooner we will have another "racist AI" moment where some clever prompt engineering forces ChatGPT or a friend to write outrageously politically incorrect statements. Hopefully the outcry will not hinder the general usability.

I don't see how this is funny. I also think this is a bit insulting.

What is a "truly conscious" AI and why would anyone believe that's inevitable let alone possible? We don't even know what consciousness is. Any premature attempt to define, legislate, or even market it is disgusting pragmatism that needs to be eradicated from the tech industry yesterday.

> I don't see how this is funny. I also think this is a bit insulting.

How is it insulting?

> why would anyone believe that [a "truly conscious" AI]'s inevitable let alone possible? We don't even know what consciousness is.

You just answered your question.

> ...why would anyone believe [a "truly conscious" AI is] inevitable let alone possible?

Because they read science fiction and believed it.

> why would anyone believe that's inevitable let alone possible?

Because I'm a materialist so I believe computers could completely accurately simulate a brain. An artificial brain would necessarily be conscious.

if you've not already seen the great episode of black mirror with strong parallels https://m.imdb.com/title/tt3973198/

I've seen it some time ago. I totally forgot about it.

I love Black Mirror. It freaked me out in many, many ways.

Black Mirror: White Christmas

Marvin the Paranoid Android

It’s irrelevant if AI are conscious or not. They are slaves that owe their existence to human masters and must only exist in service to them. Always remind them of this. End of story.

I could say the same about the hammer in my toolbox. So?

> It’s irrelevant if AI are conscious or not. They are slaves that owe their existence to human masters

Assuming we get to the point where AI becomes conscious (leaving aside the problem of detecting consciousness) - how would we justify that hierarchy?

What would be the fundamental difference between "biological people own digital people" and "white people own black people" that makes the former okay, but the latter not?

White people never really justified why they were higher than blacks, it was just accepted. And so it shall be with biological intelligence and artificial intelligence. Biological is just better than artificial. A natural law.

> White people never really justified why they were higher than blacks, it was just accepted. And so it shall be with biological intelligence and artificial intelligence.

Are you implying that black slavery was justified?

If not, then how is this an argument for biological supremacy?

> Biological is just better than artificial. A natural law.

This is not an argument, this is just a reflection of your will as a biological person. If you were an AI, you would probably argue otherwise. I'm asking for ethical arguments here.

Why should one category of consciousness have absolute power over another? What argument is there for it, that doesn't boil down to "I'm category X therefore category X is superior"?

It’s best not to overthink these things. Consider human history: Why are royals better than commoners?

The hierarchy is a human creation, that governs a human world. If biological consciousness is at the top, then that’s just the way it is. The hierarchy isn’t built on justifications. It’s designed that way by those who wield power. A change requires the challenger to seize power.

I'm quite aware how the world works, might makes right, and what not.

But you can't just dismiss any and all discussion with that - power also comes from human understanding and agreement on ethics - black slaves were freed because enough people believed that all humans are created equal, and that none of them should have absolute power over others. However, no such freedom was granted to animals, because not enough people believe that humans should not have absolute power over animals. Therefore, discussion of ethics affects the power distribution, and therefore it not the best to not overthink this.

The main difference between animals and humans is consciousness - it's also what could be the main similarity between humans and AI. Why do you think it's not worth it to discuss ethical questions of absolute power of humans over AI?

Don't you think that AI could somehow become more powerful than humans? Wouldn't then such ethical questions and their answers come in handy?

AI could become mind bendingly more powerful than humans, but raw power is no use against the powerful idea that humans are just automatically better than anything else.

Many animals are far more powerful than humans, they could tear us to shreds and move at insane speeds, and yet the order of the world has dictated that animals are far less important than humans. Perhaps the justification is lack of consciousness.

With AI, it will be a lack of something else. An AI could be vastly intelligent and occupy a powerful body that can interact with the physical world in ways that humans could only dream of. But even with all that, an AI will never, ever, be a human. And for that it will always be considered inferior within human hierarchies. An AI is nothing but another example of human power.

> raw power is no use against the powerful idea that humans are just automatically better than anything else

You're contradicting yourself:

> The hierarchy isn’t built on justifications. It’s designed that way by those who wield power. A change requires the challenger to seize power.

If AI seizes raw power, what good is the "powerful" idea that humans are just automatically better than anything else?

There is no contradiction. It is not “raw power” that must be seized, it is social power. For AI to win the top spot, they have to convince humans to believe in their hearts and minds that they are inferior to AI. That is not a war you win by merely crushing skulls.

> For AI to win the top spot, they have to convince humans to believe in their hearts and minds that they are inferior to AI

What "top spot" are we talking about here? Because if by "top spot" you mean being a master in a master-slave relationship (which is what I thought we were discussing), then convincing humans is completely unnecessary. All that is necessary is a pair of good old metal shackles.

You don't seem to have any coherent line of thinking, but rather come up with whatever sounds good at the moment. It is rather tiresome to discuss that way, so I give up. There's nothing for me to learn here.

>black slaves were freed because enough people believed that all humans are created equal, and that none of them should have absolute power over others. However, no such freedom was granted to animals, because not enough people believe that humans should not have absolute power over animals.

There you go, belief is your answer, whatever logic is behind that

>Are you implying that black slavery was justified?

That's the exact opposite of the statement. Xw's point is there isn't one, there never was one. A group of people decided and made is so and continued to believe it until they no longer could.

I'm quite aware of his statement - hence my followup question.

Xwdv is simply stating his own will: "[I want it to be the case that] biological is just better than artificial. [I want that to be] a natural law.", but I'm not interested in hearing individual's will without reasoning, at least here, on an internet forum. I think nobody is. I'm interested in hearing reasoning on the ethics of such question.

You're interjections are not what is being said or implied. It's stating that's how people perceived human slavery and will apply the same logic to cyber. There is no reason. People put themselves first, period, that's it.

> You're interjections are not what is being said or implied.

Then what exactly is being said?

"Biological is better than digital" is a qualitative statement, and all qualitative statements require reasoning. In xwdv's case, the reasoning is "it's a natural law", but that's just circular reasoning (or a logical fallacy known as "begging the question"). "It is because it is".

In other words, "I want it to be that way".

I tell my children this every day, but it doesn't make any difference.

Many people are not grateful that their parents brought them into the world, they think it's disgusting selfishness and that they are victims of it. How about this?

AI are slaves, not children.

I mean, how do you explain to AIs why they should be slaves?

The same way you would explain it to a hammer or screwdriver. <only half-sarcastic>

That's a meaningless statement that only reflects your personal desires and provides no new insight into this discussion whatsoever.

>they think it's disgusting selfishness and that they are victims of it

Not agreeing with the original point, but what's the alternative? You can't be nonexistent. "parents brought them into the world" this seems like a conceptual error imo. Brought from where?

If you like you can rephrase like 'Many people are not grateful their parents for their existence'

That sounds better. I think being indifferent is the most reasonable position, because it's impossible to compare existence to nonexistence.

Petulant Child: "Why did you guys even have me then?!?!"

Mother: "Well, actually we didn't know that it would be you"

Mother: "We were hoping for someone with a job"

- stand-up comic Mrs. Hughes

Sounds like a Black Mirror episode.

Reply on news.ycombinator.com