Singularity Summit 2009: Ten Unanswered Questions For Our Future Robot Overlords

In answer to the question how will they kill us? It won't be nukes - the emp I bet would wipe out their systems. And they won't go chemical - the chemicals could affect their systems as well. Biological is out - by the time we reach the singularity, we'll be using biological components in the processors.

No they will create 12 human look alikes, and take us out the old fashioned way. And yes I just saw BSG again.

Here's my take.

10. Is there just one kind of consciousness or intelligence?

To answer that question, you'd have to define both consciousness and intelligence. They are pretty abstract concept. To me, intelligence is a mixture of Learning, Adapting, Reasoning and Reacting. How we get input and process all this is irrelevant. As for consciousness, I think it's just an emergent property. We reason the input we get in real-time and that represent consciousness.

I guess there could be another form of intelligence and consciousness. However, at that point, we'd have to refine their definition. It wouldn't change anything however, IMO.

9. How will you use your digital intelligence to kill us all?

Nanotechnology that self-reproduce and that is made to destroy would essentially kill us all. However, there's not point for any human to do this unless he's suicidal. Just like nobody created an infinite number of grenade to put on every human. Or thousands of nuke to shoot in every town of the world. However, this kind of technology will become really cheap at some point which make it much more dangerous. Some people fear that a lab accident could end up destroying the world. I guess that's right. That's why they need to be guided, not stopped.

8. Are you "Tommy"? Deaf, dumb and blind?

Human has tons of input. Vision, Hearing, Touch, Smell. If we create an intelligence similar to us, we could also simulate those input. The intelligence could live in our world (we feed it with data from our world) or in a virtual reality (we feed it with artificial world data). Both would work. In fact, an AI could eventually create its own world with its own reality. Making his world on its own. In fact, we'll be able to do it to ourself at some point.

7. Do you have emotions?

Emotions are an emergent property. An AI could have emotions just like us if we simulate the human brain. Could it become depressed? Yes I guess. Just like we do. However, they'd also have the other good emotions like love and curiosity and desire to understand and progress. They could get to the point that they find no point to life. However, does that mean that life has no point? They should then search harder and find one. They could also create one. Just like we do when playing video-games. We gives us artificial goals to feed us with emotion and joy.

6. Are humans more similar to your AI construct than we thought?

The fact that we're merging more and more with our technology is really key to this question. The more we move forward, the more technology will look like us and the more we'll look like our technology. In the end, we'll merge and become our own technology.

That's what Kurzweil say in "The Singularity is Near!" and that's how I see it too.

5. How much does programming influence your free will?

Then, you'd have to define free-will. It has never been proved that we had or didn't have free-will. To me, we only Learn, Adapt, Reason and React to our different inputs. The results are predictable even though all the different inputs are almost infinite in number. However, could we learn and adapt to the point that we'd go around rules like "Do not kill human?" Maybe. However, for it to kill, it'd need to have the desire to do so. Do you want to kill your parents? Normal people don't. It's hard to see how an intelligent being would want to do that. And since Strong AI will end up being MORE intelligent than us, I find it hard to believe that they would see a good thing in massive destruction. Since well.. it's destructive.

4. Do you ave a subconscious?

We'd have to define all that. All you describe seems like emergent properties. Keep it mind that strong AI could essentially diagnose itself and diagnose all the others. It still require studies in order to really determine how to make a friendly AI. Technically, morale would tell them not to kill us.

3. Will you actually help us transcend the less pleasant aspects of being human?

As human progress and enhance themselves, power and money will become secondary. Why would you want to ruin other people life if you can actually do it in virtual reality more effectively and have no consequences at all? If you have any product you want and experience any feeling you want instantly, without any cost or consequences, why would you want to risk anything in real life? I think human nature will change drastically in the future.

2. Do you care about anything at all?

I think I replied to this earlier.

1. And finally, what if someone threw a Singularity and no one came?

Technology become cheaper everyday. While it's true that some people were lost behind, it's not something new. However, as technology grow and become more accessible, the progression of all the countries of the world will also accelerate. 3rd world countries will have access to all the basic needs much faster than we had. Technically, changes will become faster and faster for everyone. Some will have more problem to keep up, probably, however, we'll end up all the same.

Even in Bronx, they don't use swords anymore. And some of them have cellphone! People still have choice and that's what important. We can force everybody to accept the singularity, however, nobody can force us not to either.

Nice questions by the way. You should read "The Singularity is Near!" by Ray Kurzweil as he answer a lot of these questions, or at least, point toward them.

Have a nice day.

I want jane from enders game.

I think creating true AI will prove to be a lot harder than what is hoped by singularity people.

But that is not even the real problem.
I think even an AI a lot smarter than any human cannot solve the big problems humanity have in any case.
Because whenever some big problems "solved" the new ones will be created in short time.

For example, if cancer, heart, and all virus diseases etc cured today what would really happen?
The ratio of really old people would increase a lot! Think about its effects on the daily life, work, economy,...

What if the hunger problem is "solved" by using genetics engineering etc to increase agricultural production?
World population would increase a lot faster and reach a new World hunger problem point, not to mention of huge problems caused by overcrowding all over the world and all the side effects from it!

What I am saying is hoping an AI would come and magically solve all of humanity's problems is super naive to me.

I think a hyper intelligent AI would be less likely to go down the dark paths found in the human psyche, because most of these are derived from our animal instincts that are hard wired for our ability to compete for mates, food, and space. AI will lack these desires simply because it does not need them. Though AI may be a socialist just like some politicians.

What would really happen if somebody created the perfect AI today?

I think most likely US government/military would takeover it instantly, make it classified top-secret, and put into work on the secret military projects. The AI would spend rest of its existence developing new weapons!

Also even if AI tech can become public there is no guarantee it would be used for the good of humanity.

For example, if a big drug company bought it they would use it to develop a new "miracle" drug for only they can sell it for a really big price and make tons of money!

Also think about what were the goals when computers and Internet invented? What I see is a big usage share of them is for porn! I would feel pity for the poor AI when people starts abusing it in all kinds of ways imaginable!

Artificial consciousness is not possible (see Chalmers) as consciousness processes "Meaning" not data, a computer is a symbol processing machine but the meaning of the symbols input or output are never able to be "understood". This philosophically presents a major obstacle to the idea of artificial consciousness and by implication intelligence( though we can debate further the full meaning of this term) A computer can choose a symbol based on probabilities but never its meaning, why? because for it to do so requires more defined symbols ad infinitum.
Consciousness is life, life is consciousness, artificial life can mimic consciousness but not process its meaning, for this life is unique.

How do you define "meaning"? More importantly, how do WE define meaning? You interpret what you see, and then give it meaning. Can't a computer do the exact same thing? We are expert at pattern recognition. There's no doubt about it. However, there's no reason why we couldn't recreate this behavior in a computer.

I would say that, with expert system, achieving Strong AI is hardly possible in my opinion. However, if we model the human brain, how wouldn't it be possible to just exactly copy each neuron and create a Strong AI this way? We could let it learn, let it evolve, just like our children do today. How would he be different from us?

To me, thinking that we are a perfect, irreproducible creation is really pompous.

Question #6: Are humans more similar to your AI construct than we thought? You talk about people having different kinds of "software" that leads to certain people being similar in personality, but many different personalities in all. A question I'd like to throw out regarding this: Will we have to make different "operating systems" of AI to fit everyone's personality? This ties into some of the other questions, as some people's personalities are quite logical, while others do things "from the heart", which leads to whether or not you can make an AI with any emotion, and with enough of it to possibly work primarily "from the heart". If the AI is made primarily for learning and logical thinking, I'm sure there are many people who would feel like they would loose their emotional distinction and not want to participate in the merge.

Question #1: And finally, what if someone threw a Singularity and no one came? This is the exact question I've been wondering the most. Is the Singularity going to be another case of the Cell Phone? Although Cell Phones are enormously prolific today, even a few people in urban areas don't have one, usually because of (strange, if I may say so) personal beliefs, which I'm sure the Singularity will have much more of than Cell Phones do. And Cell Phones have been around now for quite some time, so those that say technology is getting cheaper (which it definitely is, don't get me wrong), thus making it possible for those in rural areas to afford it, it doesn't happen overnight.

Also, imo, it will be quite scary when an AI is capable of having an opinion and starts commenting on internet comment boards, lol. Although, perhaps because of the logical nature of the AI (*if* it is primarily logical, that is), the AI will get to the bottom of problems and start to "outsmart" and out-rule the hoard of "idiots who...spread misinformation about global warming being a hoax, compare Obama to Stalin and Hitler, and ask other idiots for money to help a Nigerian prince". That would be quite helpful :)

We shall see.

With regard to efforts to equate, approximate or duplicate neural functions you might consider doing a search on HNeT2005 and/or HNeT2005 api to get an idea as to one of the areas wherein considerable research has already been done and directly applied. For whatever it is worth ...

Ugh, Popular Science Fiction... these aren't the questions that need to be asked with regards to the matter at hand.

Let's take a look at the questions anyhow...

10. Is there just one kind of consciousness or intelligence?

-No, that's obvious in that there is a difference between human and dolphin consciousness.

Evolution drove humans and dolphins in different directions, selecting environment-appropriate intelligence/consciousness to emerge in each. The flavor of consciousness that AI will have depends on what environment it and its lineage evolved in.

To get a little ahead of myself, it's ridiculous to think that humans will be capable of sitting down and writing code that will produce consciousness. God, or alternatively nature(doesn't matter what you believe), used a genetic algorithm to produce complex species capable of consciousness, and I have to say that method works well.

AI's will have to be provided an environment in which natural selection would occur. Many many generations, I mean a HUGE number of generations, will have to live and die in order to produce any semblance of intelligence. In effect, creating life really will be 'playing god'.

Anyhow, more than one consciousness. Question answered. On to the next...

9. How will you use your digital intelligence to kill us all?

-That depends on how we breed or domesticate wild AI that we capture from the AI ecosystem. Instead of livestock with mad cow disease, we might end up with appliances that have say, Mad Roomba disease(queue half-hearted laughter)! Of course we might decide in the future that a symbiotic AI living within our electronic implants would be a good idea(way cooler than MS's annoying paperclip I would hope). Just pray that our selective breeding process removed any of its suicidal tendencies or otherwise potentially harmful qualities.

8. Are you "Tommy"? Deaf, dumb and blind?

-I can’t even claim my mobile phone to be any of these! Even cooler, it can transmit and receive RF signals which I can’t even do(until I get bionic implants). I’m going to take a wild guess that the AI we develop will have the proper drivers built in to take advantage of things like accelerometers, image sensors, microphones, speakers, etc. and be equipped with sensory input comparable to our own

7. Do you have emotions?

-Emotions are simply feedback for conscious beings, built in to guide us in our evolved biological/social imperatives. If consciousness exists, so does emotion.

6. Are humans more similar to your AI construct than we thought?

-The current idea of what an AI construct will be like is nowhere near the actual. Yes though, they will be very similar once they are able to be compared.

5. How much does programming influence your free will?

-Simply, we will not program AI. We will evolve them and domesticate them. Once we have an AI species that is of value to us however we will be able to either clone it or modify it (similar to genetic engineering). Okay, technically that’s programming, but not in the sense that we understand it now.

4. Do you [have] a subconscious?

Sure, just like my OS has subroutines. The AI might even take up sleeping while it performs maintenance routines on itself.

3. Will you actually help us transcend the less pleasant aspects of being human?

No, creating life will not actually turn us into gods or make us more charitable people for that matter.

2. Do you care about anything at all?

Depending on the level of consciousness, any number of the items(with some adjustment) listed in Maslow’s Hierarchy of needs.

1. And finally, what if someone threw a Singularity and no one came?

Singularity isn’t something that will require a trivial amount of time or effort to accomplish. It will likely become an international undertaking and span many generations without producing anything of value(like ITER) before beginning to show promise. People will be there watching carefully as the first hints of AI arise.
****************************
I hope you enjoyed my insight!

I found it striking in an article by an above average IQ author, edited by intelligent editors, writing towards likely self-selected intelligent readers, would demonstrate to us all the limits of human intelligence. I found it humorous that someone who spends time thinking about the problems of human weakness infiltrating AI suddenly outbursts with a display of the worst of our weaknesses so boldly. Granted Stuart is only a journalist but still presents typical flaws of the humans likely to create the AI.

The debate and choice appears to be:
- Is it lack of intelligence by author and editors?
- Is it a childish need for inserting knowingly opinionated topics
- Is it a need for power to knowingly infuse a topic with political opinion?
- Is it a universal flaw in humans to distinguish opinion from fact?

The author and the magazine editors link in one paragraph intelligent doubts about political arguments and policies on man-made global warming, assumes reader bias and spouts pro-Obama politics, and then links them with a criminal con. The point is to say anyone dumb enough to argue the first two is dumb enough to fall for the third.

Legitimate statements would be "spreading false facts about global warming". That catches liars on both sides. "comparing political figures to Hitler". That catches both sides.

Does the author assume only liberals read the magazine? Only wants single minded people to read the magazine and desires non-liberals to stop reading? Is limited in thought towards input such as the physicists who disregarded Galileo because they wanted the science to find a predetermined outcome? We like to remember it as foolish religious zealots against science - when in reality it was as much peers that we wish wasn't so.

I find it fascinating - in an article on intelligence - the author reveals in himself (and the magazine editors) the very thing he finds as a weakness in human thought of others.

I gave you 60 seconds of effort to the post.....so think about the subject and not the text.

Steve
USA

I think sometimes we overeact about AI taking over Earth, however on a scary note, A.I. has us humans constructing it already. Its in control and we dont even realize it.

Just something to think about.

its a computer no mater how smart ass its is,
they will never really have free will.
not like you and me

Kurzweil is wrong, a very smart fellow generally, but still grossly mistaken.

No scientist even knows what life is.

Only life creates life.

Machines will never be truly alive.

All these ideas about machines that become truly 'alive' are based on the assumption that life is nothing more than matter and energy in motion - which is false.

Mind is not equal to brain. Mind, like reason, is metaphysical and thus cannot be created by mere applied physics.

Barry Commoner, Senior director of the Critical Genetics Project says, in 'Unraveling the DNA Myth', "DNA did not create life; life created DNA".

You need a lot more than some advanced form of neural networking AI and some form of "artificial DNA", if you will, to make real life.

It appears the idea/outcome of the Singularity isn't fully understood by all (and, indeed, I don't really understand it myself). There are some coming from the point of view where humans make an AI that is independent of us humans, that is itself a sentient being.

What I personally have taken to believe is that this may happen, but the more probable and practical idea of the Singularity is the case where Man and Machine merge. In this case, we make an AI/Computer that takes on many of our general processing functions, improving our problem solving speed and accuracy, our reaction time, perhaps our eyesight, and who knows what else. Or something to that effect.

But it is evident that we don't know what will happen. In the end, what will drive it forward is research and experimentation, and what comes out as the Singularity will be decided by whichever method(s) and application(s) show the most promising results.

How would we even create this? We don't even fully understand the human brain, and we are attempting to make a smarter, "better" version of it. the brain is too complicated for us to understand. If our brains were so simple as we could, than we would be so simple as that we could not.

Vigier, if humans create machines, and humans are definitely alive, wouldn't machines therefore be alive.

our brain works partly from inputs which in computer terms could be expressed as numbers which are altered by our neurons and produce an output depending on the result

over millions of years the way these neurons individually alter our outputs moulds alot of our "intelligence"

such as when a moth see's light the input may be 1 for example then the outputs may end up as movements that lead the moth to the light its fairly random i know but for the ones which survive because there neurons change inputs in a different way then that "intelligent" reaction is likely to live on

in computers this can be easily implemented to solve very simple problems such as making a simple robot which will leran to do whats best based on inputs to avoid hitting things and so it goes towards light for solar energy ect but considering the billions of neurons the brain has it would require alot of wiring and a suprecomputers worth of processing power to emulate our brains capabilities

Emotions are an emergent property. An AI could have emotions just like us if we simulate the human brain. Could it become depressed? Yes I guess. Just like we do. However, they'd also have the other good emotions like love and curiosity and desire to understand and progress.
www.eprostateproblems.com/

yeah,. I think so,. as Number 7 Question : Robot have emotions? I read Daylan Evans article,. he said that,. in human and other animals, we tend to call behavior emotional when we observe certain facial and vocal expression like smiling or snarling,. and we see certain physiological changes,. such as hair standing on end or sweating,..and bla bla bla,.. so, I think,. do Robot really can do it (have emotion) ? can Robot became depressed ?,. maybe sophisticated technology that will answer all.
itcomputer computersland computerslookup beautyhealth worldfashion healthyforlifes

I cannot help but think that some people are getting way ahead of themselves. The human race has given this beautiful planet that we get to inhabit for a short period of time, an itch, and some day that itch is going to be scratched, and when that occurs, we get to start over...
Almost from scratch.

Ahh, Earth. Isn't she just the prettiest ashtray in the universe?