The ethics of AI

That ALICE thing is terrible. I just confused the entire program a second ago. It's not altogether too hard.

 

Log in or Sign up to hide all adverts.

If your after a bunch of AI'S, Bots, Agents and Data mining spiders then check out:

Http://www.botspot.com

Or failing that you want a better definition of AI:

http://webopedia.internet.com/TERM/a/artificial_intelligence.html

 

Log in or Sign up to hide all adverts.

I totally lost interest in this thread now.

Please Register or Log in to view the hidden image!

 

Log in or Sign up to hide all adverts.

(Idiot post by someone jumping into the middle of the debate)
Are we looking to give machines full sentience?
I can see the benefits of allowing a robot to make some judgements for itself. This would allow for quadrapedal and bipedal robots to correct themselves walking, and to explore dangerous areas much as we are planning to use mice, except that the mice would be able to have the strength of a robot, and could be used to rescue survivors and help with all sorts of cleanup/rescue/aid attempts without the need for radio control.
of course in that vein, the robots would be created and artificial animals.

Why would it lead to anything different than what it would lead to in a human? If we're going to create something sentient, then it certanly can learn. I find it difficult to believe that in a Philosophy forum board no one has mentioned that any sentient creature is going to have to have some sort of long and meaningful moral education that most (if not all of us) have received over the years. the reason why an AI that is not being abused would not go on a multi-state killing spree is the same reason why I or none on this board have.

Frankly, unless the robot in question has some sort of weaponry inherent in the machine itself, I as a trained chemist could most likely do the same amount of damage in the same amount of time as any robot could, if I decided to put my mind to causing death, panic and mayhem in the world. The same as any scientist could. Given a couple weeks, raw materials and a functioning blast shield, I'd likely be able to produce a battery of rockets that could at least take down a building, and i think that any other chemist could do the same thing.

Is it? Please explain your reasoning.

 

Don't compare it with real people for now. Todays A.I. is not yet as perfect as you wish. But that's such a good progress. Most of us can't even imagine how to make it 20 years ago. A.I. still requires much more time to grow. I think it is too premature to worry about AI's domination...

Please Register or Log in to view the hidden image!

 

Ismu,

Your perspective is unfortunately very common but understandable.

AI has yet to make a big impression because the raw processing power still isn’t available to allow it make decisions in a reasonable amount of time, i.e. current AI isn’t practical. Also the software is more complex than anyone suspected but the testing of new algorithms is also severely hampered by relatively slow hardware.

Please note very carefully that processing power is doubling at least every 18 months and now closer to 12 months. Intel has a 50-year roadmap with tentative solutions to the major bottlenecks that lie ahead. The increase in power is a geometric rate and not linear, and this means that while progress seems slow now it will become much faster very quickly.

The fastest Intel chip is now 4GHz, it was 2GHz until recently, and most PCs are now selling in the 1GHz range. And this progression has been holding up well since 1943 when the first computer was built (Colossus) and used to break the German codes.

In 8 years time (2010) assuming effective power continues to increase as I have described then we might (high credible possibility) see a 1000GHz chip. Imagine a PC 1000 times the power of your current machine.

You may not see these actual clock speeds since the current direction that Intel and AMD are taking is to increase the level of parallelism on a chip while keeping the clock speed the same or just a little higher. The effect is the same since more complex programs can split their activities into tasks that can be completed in parallel rather than a single monolithic sequence. And the human brain is a stupendous example of a massively parallel processing system.

So I have little doubt that very realistic and practical AI will be with us within the decade. And we need to prepare for it.

So Asguard’s topic on AI ethics is quite relevant and is going to be a real issue.

Cris

 

Riomacleod,

Isn’t it inevitable? We will be so curious to see if it is possible that it will not be possible to stop us.

Yes I agree.

Ethics: the discipline dealing with what is good and bad and with moral duty and obligation.

We share this planet with other living things, animals etc, that depend on the planetary environment for their survival. Have we ever had the right to claim the planet as our own simply because of higher intelligence? Or in fact do any other animals have any rights? What is to say that if all humans left the planet that evolution might not continue and favor some other animal and perhaps one day the dolphin might evolve superior intelligence? We should not be so arrogant as to assume we have such rights to dominate simply because we can.

My point is that once we have no further dependence on the planet then we should allow it and its dependent creatures to evolve in the random fashion that caused our existence.

But do we have a moral duty and obligation to other creatures and life? I think we should at least allow them their freedom to grow and evolve, I think we at least owe them that.

Once we venture out into the galaxy and perhaps other galaxies then I hope we will meet many other and varied life forms. Some will be less able than us and some more able. But from everything humanity has learnt I hope that life and freedom are the the most important and that we should respect these values wherever we go.

Cris

 

Merlijn,

Why? Are you sure? You seem interested inough to post the statement that you aren't intrested, and that is quite curious.

Cris

 

Asguard,

There won’t be a difference for very long.

And from Pollux –

Well not everyone seeks financial profit, that is simply the current economic mechanism that is driving progress, I hope there will be a more generally attractive goal in life soon.

But so what will be the purpose of building sentient machines?

Surely to serve us in the same way that we build any other machines. We use machines because they enhance our capabilities or just make life more enjoyable. So why would intelligent machines be any different?

I think there will be machines that have a degree of intelligence that will be below self-awareness and probably incapable of learning new tasks beyond its programming. This will be more like a ‘simplistic rules’ based program and should never really be considered intelligent, it will just appear that way. But for complex tasks, such as driving a car where good sensory input and rapid reflexes are useful and where intelligence is needed to make judgment decisions then such a degree of intelligent machine would be useful.

But we will build machines that have no limitations and where their ability to learn and adapt will be open ended, much like humans. If it has the same degree of neural network complexity as humans then it should achieve similar levels of awareness and desire as any human. In short it will be our equal. So why would we build such a machine?

1. Because we can. The same reason people decide to climb Everest, logically pointless, but emotionally rewarding.

2. To replace human workers so people no longer have to work for a living.

3. To use in environments where humans cannot venture.

4. To fight our wars for us.

5. To be our servants.

6. Companionship.

7. Instant family.

I suspect that if they are our equals then we would have no way to force them to do items 2 through 5.

Unless we can control or enslave them then their existence appears to have little benefit apart from representing a potential threat to us.

But I have said before that the rapid change in technology means they would not remain our equals for very long. The whole period of ethical concerns and how we treat mechanical equals will be very short.

So as Polux has asked – where is the profit for us?

We will be faced with a unique position; we will have created beings superior to us in every way. We will no longer be the dominant intelligence on the planet. Will we be able to control this creation? No.

If you are faced with a group of apes, who are among the most intelligent animals on the planet, then you know that it will not be too difficult to outwit them. In the same way our intelligent machine will likely be able to easily outwit us and seek its own independent survival especially if it feels we represent a threat.

I strongly believe that the only real profit for us will be to use the knowledge of building such machines and the associated technology, is to enhance our own capabilities. Uploading is the only real profit in the whole venture, i.e. to make ourselves equal to the machines.

Every machine we have ever built is because it helps us in some way. We never build machines so they can simply exist. Building intelligent machines leads us into a different paradigm where the only way we can really benefit is to become the machine.

So if we can adapt ourselves to use the same technology then we can be equals with other intelligent machines. At this point AI machines and uploaded humans will be impossible to tell apart.

At which point the question of ethics will disappear, since if we create more AI machines then that will simply be our new form of reproduction.

A point about the term AI: Intelligence cannot be artificial. Something is either intelligent or it isn’t. If intelligence exists then it is real. The term ‘artificial’ here means man made. But how it comes into existence is largely irrelevant.

Cris

 

Concider this

Cris,

I respect what you are doing, but why would anyone really want to live forever?

In a young body and able to do everything. Over the summer, i get really bored fast because I have done most things there is to do. I would imagine that after 1000 years a person would have done EVERYTHING ther is to do. Sucide/Death is the only thing left. Also it would only be remotly important to athiest, and people who don't believe in the great heaven!

Also what would happen if you uploaded your brain into something. Would you mean all the information, or the actual brain itself?

If you did the former, you would have just created two identical clones. With the same experiences DNA wouldn't matter at this point/ age. What would talking from the Human version of you, to the Robot version of you? Crazy... I just had a wierd thought what would your life line look like?------<====

If you are talking about the second one then you would be able to tell a difference in the people.

 

Hm. Interesting. Kinda scary to think what would happen if some machine AI unit got outta control due to a computer bug and gained consciousness. Ouch.