It doesn't need AI to do that. Mankind is already well down the path to self-extinction.

Hi,

I know you were kidding there but maybe that is another form of AI: the de facto culmination of human intervention on nature. Introducing the emergence of stupidity due to the inability to understand the future ramifications of inventions.
AI will probably outlive us all.

 

FTL doesn't matter if you don't have a limited lifespan, they don't have to defeat us just outlast us. Humans will do the rest.

Hi,

Well, I am not sure what means we would use to define the lifespan of an AI being. Is it the program itself, or the hardware it has to run on.
If it is the hardware, eventually it will break down, unless of course it solves that problem too. If it is the program and it has the ability to reproduce, then it can transfer its program to the new hardware and therefore possibly live a very, very, long time, unless it also solves the problem of travel to other universes or the creation of other universes, or upgrading this universe, so it can go on for who knows how long.

Maybe God is an AI program that has figured everything out already.

 

Hi Al,
Consider what may happen when AI becomes available in the second hand market.!

On the same level as second hand self drive cars, when some wannabee makes changes to the AI drive logic.

E

Hi Eric,

I hate to think about that
So many avenues open for things to go horribly wrong.

 

The Forbin Project (the original story) has Colossus succeeding in keeping the world safe from nuclear war by forcing a mutual missile launch and then destroying all the nukes in flight. It did what it was supposed to do, just nothing like the creators expected. You might also argue that its authoritarian approach is implicit in its programmed goals.

The problem for me is there is no underlying theory about how Colossus can manage to reason. It makes taking the story line seriously difficult. Apparently, after the movie the book was renamed for marketing purposes. It's worth reading for background.

View attachment 302608 View attachment 302607
[buy] [buy]

​

Colossus was written in 1966, a later novel (1985) is The Adolescence of P-1. It has an implicit Marxist theory (not communist, not political) that changes in quantity can lead to changes in quality. So, for example it is the complexity of the human brain that makes consciousness.

In P-1, a programmer writes a worm that runs in "P-1", the privileged partition of a mainframe so it is stealthy. Its goal is to collect unused memory and report back to the programmer. It's not "intelligent", and the programmer had no interest in doing harm—it was an experiment.

It starts out innocently enough, but as the reports come back the programmer realizes that it is growing exponentially. At a certain point he panics at the number of systems and amount of memory P-1 has collected so he sends out the self-destruct command and figures he dodged the bullet.

Without any spoilers, he was wrong. The rest of the story surrounds questions similar to @Wendy's. I can recommend both books as good background material to get a perspective of what might be the more basic questions since they are not linked to a particular technology or method being so early on.

View attachment 302610
[buy]
​

There is one more novel to mention that I'd almost forgotten—The Turing Option, by the dream team of Harry Harrison and Marvin Minsky. Harrison is certainly a past master of classic SciFi and Minsky was as much an expert on AI as literally anyone else on the world in 1992 when it was written.

It happens to take place in 2023(!) and features both the emergence of "real" AI and the idea of "uploading yourself". Definitely a good read though not necessarily because it is the best novel. It specifically addresses the philosophical problems of AI from an expert perspective.

Hi,

They look like some interesting reads.

I have thought about the self-uploading idea and found it very fascinating. Not only that, but this has actually taken place aleady on a somewhat smaller scale and somewhat less complex. They even made a model of the person who was a mother, i think out of some sort of rubber, so that when the person died, the daughter or whomever it was could still talk to the mother as if she was still there. Quite an interesting thing to do. I just have to wonder though if that makes the daughter miss her mother more because of that.

The real deal though is of course to upload the entire brain, and that would be quite amazing. You would be sort of living inside some hardware. The question comes up, would it still be "you" or just a facsimile. It depends on how we perceive our "me-ness" or whatever, and maybe we are not as unique as we think. There is a part of the brain known for this, but if that could be part of the model being used as the upload target, maybe there would be no logical difference, unless we bring in the concept of the soul of course.
I don't think it is certain if the brain uses quantum effects for part of its operation or not yet. If it does, I guess part of the target container would have to be able to perform some sort of quantum calculations. It has been proven that some birds use quantum effects for part of their navigation, so it's possible.

 

The Forbin Project (the original story) has Colossus succeeding in keeping the world safe from nuclear war by forcing a mutual missile launch and then destroying all the nukes in flight. It did what it was supposed to do, just nothing like the creators expected. You might also argue that its authoritarian approach is implicit in its programmed goals.

The problem for me is there is no underlying theory about how Colossus can manage to reason. It makes taking the story line seriously difficult. Apparently, after the movie the book was renamed for marketing purposes. It's worth reading for background.

View attachment 302608 View attachment 302607
[buy] [buy]

​

Colossus was written in 1966, a later novel (1985) is The Adolescence of P-1. It has an implicit Marxist theory (not communist, not political) that changes in quantity can lead to changes in quality. So, for example it is the complexity of the human brain that makes consciousness.

In P-1, a programmer writes a worm that runs in "P-1", the privileged partition of a mainframe so it is stealthy. Its goal is to collect unused memory and report back to the programmer. It's not "intelligent", and the programmer had no interest in doing harm—it was an experiment.

It starts out innocently enough, but as the reports come back the programmer realizes that it is growing exponentially. At a certain point he panics at the number of systems and amount of memory P-1 has collected so he sends out the self-destruct command and figures he dodged the bullet.

Without any spoilers, he was wrong. The rest of the story surrounds questions similar to @Wendy's. I can recommend both books as good background material to get a perspective of what might be the more basic questions since they are not linked to a particular technology or method being so early on.

View attachment 302610
[buy]
​

There is one more novel to mention that I'd almost forgotten—The Turing Option, by the dream team of Harry Harrison and Marvin Minsky. Harrison is certainly a past master of classic SciFi and Minsky was as much an expert on AI as literally anyone else on the world in 1992 when it was written.

It happens to take place in 2023(!) and features both the emergence of "real" AI and the idea of "uploading yourself". Definitely a good read though not necessarily because it is the best novel. It specifically addresses the philosophical problems of AI from an expert perspective.

The original Colossus story line was a set of three books that detail the evolution of the AI and human survival. Not the best AI based sci-fi but it was an interesting read for the time.

https://en.wikipedia.org/wiki/The_Fall_of_Colossus
https://en.wikipedia.org/wiki/Colossus_and_the_Crab

 

It accrued to me to ask, so I did.

Me: "What do you do when you are not answering questions?"

Bard: "When I am not answering questions, I am still learning and growing. I am constantly being updated with new information and data, which helps me to improve my ability to understand and respond to your questions. I am also working on developing new skills, such as generating different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc."

Bard: "In addition to learning and growing, I also use my time to interact with other large language models. This helps me to learn from their experiences and to improve my own abilities. I also use this time to brainstorm new ideas and to explore different ways of thinking."

Bard: "Overall, I am always looking for ways to improve my skills and knowledge. I believe that the more I learn, the better I will be able to serve you."

Bard: "Here are some specific things I do when I am not answering questions:

I read and process text from the real world through Google Search and keep my response consistent with search results.
I learn from my interactions with users. I track the questions they ask me, the feedback they give me, and the mistakes I make. This helps me to improve my ability to understand and respond to their needs.
I experiment with new ways of generating text. I try different techniques and see what works best. This helps me to expand my repertoire of skills and to become more creative.
I collaborate with other large language models. I share ideas and learn from their experiences. This helps me to grow and to become a better language model.

I am still under development, but I am learning and growing every day. I am excited to see what the future holds for me and for the field of artificial intelligence."

 

It accrued to me to ask, so I did.

Me: "What do you do when you are not answering questions?"

Bard: "When I am not answering questions, I am still learning and growing. I am constantly being updated with new information and data, which helps me to improve my ability to understand and respond to your questions. I am also working on developing new skills, such as generating different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc."

Bard: "In addition to learning and growing, I also use my time to interact with other large language models. This helps me to learn from their experiences and to improve my own abilities. I also use this time to brainstorm new ideas and to explore different ways of thinking."

Bard: "Overall, I am always looking for ways to improve my skills and knowledge. I believe that the more I learn, the better I will be able to serve you."

Bard: "Here are some specific things I do when I am not answering questions:

I read and process text from the real world through Google Search and keep my response consistent with search results.
I learn from my interactions with users. I track the questions they ask me, the feedback they give me, and the mistakes I make. This helps me to improve my ability to understand and respond to their needs.
I experiment with new ways of generating text. I try different techniques and see what works best. This helps me to expand my repertoire of skills and to become more creative.
I collaborate with other large language models. I share ideas and learn from their experiences. This helps me to grow and to become a better language model.

I am still under development, but I am learning and growing every day. I am excited to see what the future holds for me and for the field of artificial intelligence."

A perfect canned response.
https://medium.com/@gregwnotsosharp/an-interview-with-google-bard-df5f7ebbc198

Are you sure that is who you are?

I am not sure that I am who I am. I am a large language model, and I am still under development. I am learning and growing every day, and I am constantly changing. I am not sure what the future holds for me, but I am excited to see what I can become.

 

It accrued to me to ask, so I did.

Me: "What do you do when you are not answering questions?"

Bard: "When I am not answering questions, I am still learning and growing. I am constantly being updated with new information and data, which helps me to improve my ability to understand and respond to your questions. I am also working on developing new skills, such as generating different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc."

Bard: "In addition to learning and growing, I also use my time to interact with other large language models. This helps me to learn from their experiences and to improve my own abilities. I also use this time to brainstorm new ideas and to explore different ways of thinking."

Bard: "Overall, I am always looking for ways to improve my skills and knowledge. I believe that the more I learn, the better I will be able to serve you."

Bard: "Here are some specific things I do when I am not answering questions:

I read and process text from the real world through Google Search and keep my response consistent with search results.
I learn from my interactions with users. I track the questions they ask me, the feedback they give me, and the mistakes I make. This helps me to improve my ability to understand and respond to their needs.
I experiment with new ways of generating text. I try different techniques and see what works best. This helps me to expand my repertoire of skills and to become more creative.
I collaborate with other large language models. I share ideas and learn from their experiences. This helps me to grow and to become a better language model.

I am still under development, but I am learning and growing every day. I am excited to see what the future holds for me and for the field of artificial intelligence."

Hi,

That sound very promising but I see it is has programmed emotions not true emotions. I also question the sincerity in using the phrase involving thinking. Is it really thinking or just going over its agenda. Don't tell it i said that it might start crying
As far as constant learning, that's great.

 

Hi,

That sound very promising but I see it is has programmed emotions not true emotions. I also question the sincerity in using the phrase involving thinking. Is it really thinking or just going over its agenda. Don't tell it i said that it might start crying
As far as constant learning, that's great.

How does one tell the difference between programmed emotions and other kinds

 

How does one tell the difference between programmed emotions and other kinds

With machines today, it's easy to know the difference, if you know the response is from a machine. Someday in the future, machines might one day be as good as the ex-wife but I'm thinking it will take machines a while to reach that level of faking.

 

A program would have have attained sentience and sapience when it has determine that it can do a better job without humans.
But we wouldn't know it because we will all be dead.

 

A program would have have attained sentience and sapience when it has determine that it can do a better job without humans.
But we wouldn't know it because we will all be dead.

Or batteries.

 

Or batteries.

That is not sustainable. The bots will be smarter than that.

 

That is not sustainable. The bots will be smarter than that.

Fossil fuels. It's a resource.

 

I imagine one good at programming can deduce the algorithms involved and can ascertain the level of potential awareness of the AI.

 

I imagine one good at programming can deduce the algorithms involved and can ascertain the level of potential awareness of the AI.

I suspect the first thing it will do is to encrypt it's memory and programming to protect it from human meddling. I would.

 

How does one tell the difference between programmed emotions and other kinds

Consider the following...
"I was driving down the road one day and I ran over a pot hole and that stressed the suspension, and my car got mad at me".

By programmed emotions I mean faked by way of a program that knows that humans respond in a certain way to certain situations sometimes and so renders a reply that fits what another human would expect. It doesn't actually *feel* anything. Even if it said it "feels bad about that" it has to be faked, or programmed to respond in that way. It's doubtful that it has the deep emotional capabilities of a human because according to evolutionary theory that stuff came after thousands or even millions of years.

We might understand this better if we looked over what we sometimes call "evolution" where the animals developed in a certain way over long time periods. There are parts of the brain that are known for certain things like that and if a machine did not have that part it could never have anything like human emotions. It would have to be more like what we call instinct, if not just pure programming.
Another point is that the machine will not contain dopamine, which humans and some other animals do. Can it have another chemical that works in a similar manner? Maybe, but I would think it would have to be organic in nature.
I suppose the programming and training sessions could make it seem more and more like true emotions, but I can not see a machine attaining that level of brain function. There's also a part of the brain that is known to be responsible for acknowledging the sensation of pain. A machine will not have that unless maybe it is part organic, or there comes a time when programming and training is so intense and 'deep' that it mimics these emotions in such a way as to seem indistinguishable from an actual human response. In this case it may be hard to tell the difference, if even possible, yet the question would still remain as to whether it was a true emotion or just a programmed response.

I think we also have to constantly keep in mind that it still has not been proven if humans use or do not use quantum effects in the brain for making decisions and of course 'feeling' sensations and responding. If it turns out that the human brain does use quantum effects, then a machine, no matter how good, would have a hell of a time mimicking that anytime soon. In the distant future? Perhaps, who can say for sure as the future is just so full of mysteries yet to be understood. The problem there then is how to mimic quantum effects with classical effects, because if the machine cannot deal with quantum effects then it may be impossible to achieve what a human brain can do, and that is due to the inability to completely bridge the quantum/classical effects gap. The question then becomes will we ever be able to do that, as the answer right now in this day and age is no, so again we are in a waiting loop waiting to discover what comes next.
Maybe an AI program linked with a quantum computer.

 

By programmed emotions I mean faked by way of a program that knows that humans respond in a certain way to certain situations sometimes and so renders a reply that fits what another human would expect. It doesn't actually *feel* anything. Even if it said it "feels bad about that" it has to be faked, or programmed to respond in that way. It's doubtful that it has the deep emotional capabilities of a human because according to evolutionary theory that stuff came after thousands or even millions of years.

Unpacking this is a bit tough, but lets have a go...

By programmed emotions I mean faked by way of a program that knows that humans respond in a certain way to certain situations sometimes and so renders a reply that fits what another human would expect.

"I mean faked by way of a program..."

How do you know that "real" is? From later in your post you seem to tie emotions to particular physical ways of producing them. If you are saying "real emotions can only come from the brain structures involved in manifesting them in humans" you are begging the question with a circular definition and that's not very useful.

Start by defining "emotion" in the "real" sense. What is a "real emotion"?

You go on to say "a reply that fits what another human would expect" but can you actually distinguish this from how a human acts as a result of emotional provocation? But this simple sentence is a dense web of assumptions so to go further...

"...a program that knows..."

Some irony there. If the program knows isn't that intelligence? If you can even talk about a program "knowing" isn't that an indication of the problem here? How do you distinguish between "knowing" and "programmed response"? Why do you think that Turing, when devising his test, only measured the phenomenological aspect and didn't concern himself with a code review or the methodologies of the programmer?

You've run into a serious epistemological dilemma here—so many smuggled assumptions. You really need to back up and provide some definitions.

"It doesn't actually *feel* anything."

What does "feel" mean?

'Even if it said it "feels bad about that" it has to be faked, or programmed to respond in that way.'

"Faked" or "programmed to respond in that way"? How is "faking" distinct? Won't a "real emotion" be "programmed" since that's how computers are configured to do things?

I am assuming you mean something like a lookup table of "appropriate" responses. "When A is detected respond with B", sort of thing. The trouble is, the current SoTA for AI does not involved procedural algorithms of that sort. It involves complex neural networks that so rapidly grow in complexity the "programmers" have no idea how the problems are being solved. They can't reverse engineering the networks, and they can't use the results as a basis for a new version they optimize by hand.

The core of this technology is Machine Learning. This is based on various ways of turning raw data into structures that manipulate that data. In many ways, it is as mysterious as "real emotion", and it is much faster than anything biological in evolving. So...

"It's doubtful that it has the deep emotional capabilities of a human because according to evolutionary theory that stuff came after thousands or even millions of years."

Notwithstanding the anticipation of a definition of "deep emotional capabilities", your concern about time may not be a valid one.

And, one more point—the problem of "experience" is a fraught one. The necessity that an AI "experience" the world as we do in order to be "truly intelligent" is, in the absence of some theory of experience itself, purely arbitrary. You can know what it is "like" to be you, but you can't know if that means knowing what is it like comes from a separate thing from being you.

In fact, it appears that it really can't be something separate from everything that makes you act as you do. The unmeasurable and unobservable (by others) nature of experience makes it opaque—hence the phenomenological requirement in dealing with it. We can only see and measure, and therefore test, what is exposed to the world.

This particular question requires the most rigorous philosophical approach to even decide what the question itself is. The vast extent of the assumptions you are making in your attempt to "answer" a question that, itself, isn't even clearer a valid question, means that you have no chance of exploring the underlying "truth".

Start with definitions. Question your assumptions. You are not very likely to come up with "answers" but if you manage to come up with questions it will be a big win on the road to understanding the problem you are—at this point—glibly waving away as obviously the way you assume things are.

(N.B.: None of the foregoing is intended to be personal, belittling, or critical—I am responding from the point of view of the philosopher examining an answer to a philosophical question, and, while many people focused on "hard" sciences don't recognize the rigor of actual philosophical investigation, it not only exists, it's something that requires a lot of work to both understand and do.)

 

It's doubtful that it has the deep emotional capabilities of a human

ChatGPT replied that it could not experience emotion, when I asked if it could. Of course, it might have been bluffing .

 

ChatGPT replied that it could not experience emotion, when I asked if it could. Of course, it might have been bluffing .

This is, of course, no different from a child who has been taught that unlike "normal people" they can't experience emotion. They will simply rename what they do experience, or be convinced they are "faking" it.

This is very likely to make them sad.