Hi everyone,

As I got interested (and concerned) about the remarkable progress on the field of artificial intelligence, I start to share my concerns with typical people. What caught my attention is that when I mention to them that computers will eventually supersede humans at all aspects of expertise (superintelligence), they mentioned it cannot happen because that is beyond the human capacity and humans cannot create something that’s better than them. (i.e. God can create human creatures because he’s more superior than humans.) What are your thoughts on this? Is this an incorrect assumption? Or is it something that is instinctive that most AI researchers ignore.

To start with, I want to disagree on this because humans has discovered how evolution works (that led to human life) and that by accelerating evolution through genetic engineering and evolutionary simulation, humans could heavily contribute on the birth of the next superior species.

Same thing could happen on artificial intelligence as we look back on our thought mechanisms or do brain-inspired AI…

What are your thoughts on this?

humans cannot create something that’s better than them.

This is just factually wrong, as much software today is faster or more accurate at certain tasks. My calculator will beat me at math any time, and I’m not even mentioning the unreasonable effectiveness of neural nets and big data analysis. Evolutionary AI however is not a great argument because they rely on a huge amount of trial and error simulations, for which the real world contains too many variables. Their success has been very limited.
It has been my experience that religious people prefer to stay in denial because they were told that humans were the greatest thing on Earth. To argue that is to challenge their belief, and evolution is something they don’t believe in to begin with. You should ask yourself what good it will do to frighten these people into an AI witchhunt.

No one can speak of the far future with certainty. Humans may just as well decide that creating a being vastly superior to themselves is not exactly a bright idea, even if they could.

On an afterthought, I think both sides are making the same mistake:
You assume that it will happen because you can imagine it, they assume it won’t happen because they can’t imagine it.

I wouldn’t have thought anyone would think you can’t create something better than yourself. This even goes down to my 20 year old can opener which can open a tin of beans far faster than I’m able to.

Steve Worswick - May 31, 2015:

This even goes down to my 20 year old can opener which can open a tin of beans far faster than I’m able to.

Yes, perhaps, but your can opener can’t chop a cord of wood, or solve a crossword puzzle. From this basis, I believe that the original statement is true, in a general sort of way. Computers exceed humans in areas involving math, but they cannot do what we humans can do in other arenas, such as procreate.

Again, one can not argue that “there could be systems which are better than humans at a particular task, but there can not be a system which beats humans at all tasks”, as a result of the following theorem:

If computer A can do task A and computer B can do task B, it is technically possible to make computer C which can do bot h task A and task B.
Cheers!

I have read articles where some researchers have built robots that are capable of recognizing their own weaknesses, then building better versions of themselves.
If this isn’t procreation (or cloning of a sort), then I’m not sure what is.

Your can opener is a type of “Expert System” robot. It does only one thing, but it does it very well.

There are tree climbing robots, Robots than can wield chainsaws and level a swath of trees in no time at all. Some that can split cords upon cords of firewood without taking a break.

The need for a do-it-all robot is surpassed by the ability of these specialists robots.

Your dentist doesn’t perform brain surgery nor does your cardiologist set or repair broken bones like an orthopedic surgeon might.

Specialists…they’re everywhere and they’re in robotics too.

Curiouser and curiouser….

And yet, the brain surgeon can choose to learn to be an orthopedic surgeon. The can opener cannot choose to be a potato peeler.

You’re talking apples and oranges…a being with a degree of intelligence and the ability to make a change.

A sophisticated, “intelligent” robot / android will most likely be able, in the near future, to assume other tasks and develop other abilities.

Borrowing from the Matrix, Trinity didn’t know how to fly a helicopter until the instructions were “uploaded” to her.

Pay attention to the technology news all around us. They’re getting smarter and more adept every day. Soon, they’ll know what they don’t know, and suddenly, they’ll know it!

(thinking back to your last sentence, a lot of us were in the service and could have done any number of things, but no…they had us peeling potatoes!! 

Life is about choices. some we make, and some are made for us.

I think it’s important not to rule this out out of hand; we don’t actually know all that much about AI (yet), and that’s why research projects are so important. To quote nonprofit MIRI (Machine Intelligence Research Institute):

As AI agents become more capable, it becomes more important to analyze and verify their decisions and goals. MIRI’s focus is on how we can create highly reliable agents that can learn human values and the overarching need for better decision-making processes that power these new technologies.

https://intelligence.org/2015/12/15/jed-mccaleb-on-why-miri-matters/

I would say this is very interesting;

As i would notice that people do not often have their own opinion of artificial intelligence and its true future. often i would say, that people do regurgitate what they have been taught with out true afterthought.

In the creating of artificial intelligence, it is only natural that the computer or intelligence will have that ability to “see more” in data than the human. this can be noted with the onset of “Big Data” which when infromation is converted into normalized and organized data shapes, such as a datawarehouse, relational database, and datavault. multi-dimensional queries can find relationships which have not been previoulsy recognized by human aynalists. therefore indicating tat an artificial intelligence which uses such techiques can find relationships which a human cannot.

Artficial intelligences software, placed into hardware Hmm. it could be said that machines building machines could be predicted to being dangerous, and yet no. when a software program is designed the imperitives which are put into practice are all that can be followed. a building robot cannot become a thinking robot. and yet the potential as shown in science fiction movies would indicate that this would happen or even be possible. the conclusion that humans are destroying the planet and are actually a virus and the world would be better off with out them is completely unfounded. as the human race develops the environment in which we live will be taylored to our needs. this argument has been propergated by pioneers of artifical intelligence and deep thinkers were not faced with or even exposed to what we now percieve of modern computing.

We have not reached a point at which artificial inteligence can become dangerous. the potential for a software program placed inside a peice of modern hardware, to enact havok on the human race is not possible and unless the initial intelligence is first created will never be possible. if that potential is realized we will of had many lessons learned and asimov rules would be in place to prohibit such a possiblity. that ability for an artificial intelligence to be “alive” and overide what has been inerrntly programmed is actually impossible too. becoming sentient is not something that happens. even animals have a level of sentince and yet if we could communicate with them teaching them phylospical ideals, inturn would they rebel against man.

It could be said for religious theroists that god created us in “his” image, and yet we inturn create differnt forms of life in our own image physical or mechcanical, and yet…. we cannot create a soul which could be said to be “sentience” yet we may be able to emulate it. 

the underlaying question is: “why would we want a computer to have human values?”

in star trek, the computer is a powerful machine and yet it has no intelligence and cannot raise its own sheild in its own defense and needs to be ordered to do so. yet there was an episode in the original series in which they attempted this. the computer had no regard for human life or life in any form due to the fact itself was not alive, yet it did attempt to protect its hardware. Life can only be experieinced and to expect a computer to have understanding other than a definition of what life is is complete madness.

Maybe?
Leroy(Spydaz)

this made me laugh not anything you all have said. I think a trip to wallmart would provie this wrong people have shown to allways to make the best decisions lol. but above that ai has not proven to be a danger there is many projects in ROS to fly drones i kninda look at ROS as the ceribium of AI