Artificial Intelligence at Google

[JonnyTommyGuns]

If it got bad the government would step in and control it. Beside that, Google is not a group of evil geniuses and the AI itself would be incapable of empathy or apathy. Just pure mathematical logic. It would be nice to have a super intelligent machine to help solve the world problems. Obviously it would not be in control of anything it would simply be a think tank.

[plmqas]

you are stupid man, each evil genius are able to control human variable !!! BUT THEY SIMPLY DONT GIVE A DAMN ABOUT THEIR FEELINGS and lifes

[XanaduZune]

Empathetic people often have foolish political ideas that cause serious problems. For example if you feed starving people, they're just going to breed a lot more starving people and cause a bigger problem. That's what happens when you use feelings instead of logic and objectivity. Also today's great animal species evolved through the brutal survival of the fittest principle, not the principle of helping the weak to survive. If you help the weak to survive then you weaken a species.

[adikid89]

the thing with artificial intelligence is that... Human being are not dumb.. as opposed to the God of the bible... we always have a off switch!

[MrBrian987987]

this guy is a nerd

[gerviex]

Alan M.Turing (1912-1954). Best Known As:

PIONEER IN COMPUTERS AND ARTIFICIAL INTELLIGECE.

SearchAnyway!! Thanks for this vid .. interesting.

[OcidBand]

Now if said AI bot was in fact lacking empathy, than that same bot would lack AI. AI is not only simulating one perticualr aspect of the human brain but all in it's entirety., I have meet all types of genius' in this world and I Must say quite interesting as to just how our minds work (in a sense integrating Genius and Soul).

If the AI bot was allowed the internet, I would assume as it learned it would call its self GOD....

[eeemotion]

only the respect of Asimov's three laws can save us from robot har. Unfortunately most are built by the military so the last thing they wanna do is just that.

[realguy420]

While the many basic primative versions of AI will and already do things for us that help, many or not-so-many years down the line the ultimate goal is the replacement of humans with robots or cyborg droids, but hey we'll have some nifty gadjets til then eh?

[DJDAVIDXJOVAN]

You are T-1000 you replicate human skin.

[ThinkTank255]

Nonsense. AI will think for itself, but only in a purely logical way. Without the emotional drive telling it to do harm AI is about as harmful as a teenage nerd with no life.

[snapper842002]

it doesn't have to have an emotional drive, it can be maliciously programmed to have a specific goal state which is dangerous to all or certain humans

[gerviex]

@snapper842002

I fully agree in what you say

[DanFrederiksen]

it's an understandable point and an important issue. I myself stopped working on AI for ethical/spiritual reasons.

but your point is not actually accurate. it's a classic fallacy to confuse intelligence and will, just like a car can drive fast doesn't arbitrarily decide to drive you over fast. an intelligent system will have no inherent desires of its own and certainly not any evil. if designed to kill it would do so without hesitation but it would never be fickle like humans. evil is real

[Remindor]

There are different ways to create an artificially intelligent system some safer than others. Evolutionary computation for example has been very successful, but unfortunately we have very little control over what is produced. Other methodologies such as ANN are pretty safe since you can pretty much choose the parameters to help the machine differentiate right from wrong... You could literally program empathy into the machine via assisted learning.

[wubs23]

if any company would ever be able to make AI, it would be google. think about it.. google is BY FAR the biggest database of human behaviour in the world until this point. and it keeps growing. it knows what we search, how we search it, who we ask about it, where we travel from, where we travel to, who we have contact with... and that is all that you need to know to make AI. (and ofcourse how to program it.. but google only has the data needed)

[Swamy456]

Idiots made this video have 2 1/2 stars when he was CORRECT. 5 stars man.

[Azilythe]

There's already powerful AI in existence, but it does a DDOS attack to any brave programmer willing to probe any one of it's communities.

[matthewsjc1]

Ho hum... Guy I'm a C/C++ systems developer and I can assure you that this "robotic invasion" you're so worried about isn't possible unless we specifically design/program them to do so. Notice I said "specifically". Think about it, genius ;) . matthewsjc1 out...

[defragen1]

lets just say anything that has been made to help man kind has also been used to harm mankind...

[Swamy456]

You idiot, if its sentient it can write its own code. Manipulate its own programs.

[wiek]

this guy doesn't now anything about AI or intelligence... too bad.

[Reve101]

I'm sorry if I came off a little too strong but ignorance and giving false information to the public really pisses me off.

[Reve101]

A simple pong game has an intelligent agent (not very intelligent I must admit), intelligent agents are everywhere now a days, Google map could be a simple problem solving agent by searching that utilizes a path finding algorithm (like A*), now was that really dangerous? So I suggest you do some research, read a little, and get your facts straight before you publicly humiliate yourself.

[Reve101]

hahah man you really have no idea what you're talking about, do you. I don't know what Google is up to but the only kind of agent that could pose any danger is a learning agent, and event then it's not that simple; the agent has to have a certain goal. Normal agents for problem solving, planning etc. that are simply logical pose no threat what so ever, unless you program them to do so.

[Blankname101]

I feel we humans need be upgraded , so that we can think over a million times our current IQ(Of course start with just 1,000 times and then million and up)...Also I feel war will no longer exist once we think at over a million IQ..We would have learned to give up on this thing about being different and grow.

Peace!!

[Swamy456]

If its possible to do this with the IQ we have now its ridiculous to say we need a million iq (which is a ridiculously arbitrary number and does not make sense because IQ is a proportion anyway) to make the world better.

[Blankname101]

A.I can think over a million(And they can increase this by as much as they want..) moves to take in his mind.They will be too cunning!!One day can be over 100,000 years for a A.I , so for one to fight something that power is crazy...They will see us as pests who fight eachother and kill , rape , over populate , hurt the earth and the animals who we live with , etc....

Peace!!

[jaberwocky6669]

All minds can be controlled! MUH HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HAR HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA!

[angel2901]

@ FMMAROTO

I couldnt agree more! AI will stop the idiot behavior of the humans!

[Jabulon88]

vote ron paul

[FMMAROTO]

It is precisely the "human element" that made Hitler and Stalling evil.

[acrazyelf]

we already have ai with hanson robotics and stephen thaler creativity machine were they are both completely able to think and feel for themselves this has been a fact for along time. but the main thing is to limit there capability there could be a problem yes i agree but the thing is that there is no risk with Google ai . O and he is obviously intelligent you can see this by his comparisons. o and ai will pose a risk to humanity but no in the way described. Contact me at <bikingbullet@mac.c0m>.

[angel2901]

Fool! AI will save humanity...

[vestalunch]

i don't tend to bash people online...

but this guy is a tool

[SearchAnyway]

Would you say that I'm more of a ratchet or a crescent wrench?

[kunfushuss]

you're like a hammer and the world is a quadriplegic. The world has no use for you.

[SearchAnyway]

Well, at least the world is more pitiable in your eyes than I am... I'll take that...

[kunfushuss]

However, I (and we, in the AI community) realize how important empathy, and emotion as a whole, is to intelligence (and AI). In fact, I would say that a lack of empathy may have made Hitler bad, but so did having too much resentment (another emotion that he DID have.) Emotions are all important, and to say that our robots are going to be out of control because they lack empathy is dumb. You need to stop watching movies.

[SearchAnyway]

Whatever, dude, you don't even vlog yourself, Buzz off...

[kunfushuss]

You're an idiot. First of all, you're only talking about "Strong AI" - the idea that computers will think just like humans. That is NOT what AI is all about - we do so much more (think vertical worlds - like the ability to win chess) However, the point that made me think you're an idiot is when you said "Empathy is part of intelligence" and then said "AI lacks empathy." If so, it is NOT INTELLIGENT according to your definition and therefore NOT AI.

[SearchAnyway]

I said that empathy was one of ELEVEN kinds of intelligence... Some people have more than others... AI would probably be void of it... You should listen more and talk less...

[Nick050607]

somebody has being watching the film "stealth"...

[SearchAnyway]

Honestly, I've never heard of that movie...

[leetrice12]

@Nick050607 and terminator

[kamari11694]

@leetrice12 and iRobot

[desspec]

It is misleading to suggest that the "artificial" in "artificial intelligence" has any relevant meaning here. True intelligence will likely be something far different than what we can understand at first, even if it springs from our efforts. But that doesn't mean it will be malevolent.

[kelkora]

chris, you views sound like a man stuck in the dark ages, AI intelligence won't have any empathy but neither will it have the capacity for any other states of emotional being of which your "evil intelligence" had plenty. It will be new and simple which will require a understand on par with encountering a new sentiant specie, sure we won't have complete control but why does that course so much fear.

[SearchAnyway]

I guess I'm most concerned about the likelihood that WE will f**k up when we create AI, and that it will (as a result) forget to factor in the "human" element when allotting resources or implementing a decision.

[kelkora]

You come up with a good point human error is the great danger for mankind.

[sovereign777]

The first Government or organisation prepaired to pay me gets it first.

[mistamustard]

Sample first, please.

[fldz]

Yeah, me to, gimme a ballpark idea of the theory. I'm into that in a silly A* game-AI kinda way.

[sovereign777]

Your favorite colour may be blue and mine may be red yet that information about preference to colour is stored in the same place in our independant minds. To access it our brains execute the same operation in order to retrieve our preference to colour. Understanding the commonalities of that infrastructure allows us to understand the ones that varied our preference to colour in the first place.

[fldpi]

That's interesting, I've never thought of it that way. The anatomy is going to be common. Would be interesting to get two people somehow synchronized in reactions to watch the similar patterns. God, I wish I had zillions of dollars to mess with stuff like that. (like gov's do)

[sovereign777]

I've privately worked on A.I for 20 years and I have the answer to a subservient yet intellectually superior machine,which would be of extreme value to the human population and absolutely no threat.

[pdutube]

Oh come on. Artificial intelligence means so many different things. Just because Google is pursuing their idea of artificial intelligence doesn't mean some sort of cognizance or self awareness. Most of what is considered A.I. is really just adaptive programming. In fact many researchers can't get A.I. research money because the whole idea has been so unfruitful. We have a long way to go before we get to Sci-Fi level A.I. It takes a supercomputer to beat a chess master, and that's all it can do.

[vaflf]

This guy doesn't understand AI. Period.

[glitchcore]

There is a universe of minds that could potentialy be created, the human mamalian mind occupies just a small corner of mind space. The problem of our generation will be to reach into mind space and create an AGI that is rational and friendly. We need the AI's to be our partners not our enemies. By the way, just because it is artificial does not mean it can not posses empathetic intelligence.

[thiazisagestained]

Wow, AI alarmism based on, wait for it, Hitler analogies. It may be that good dialog on the possible effects of this technology would be better served by, and this is just off the top of my head, spending more time on learning about the structure and implementation of functional AI and less on trying to save Sarah Connor.

[fldz]

I dont think you can have an autonimously goal-oriented program without already having superimposed human values into the goals. Our problem with Google is the same one as government. When Gov says 'patriot-act' it translates into the opposite. When Google's employee motto is 'dont be evil', they mean 'dont appear to be evil, dispite your true nature'. Would that our future was entrusted to pure logic instead.

[SearchAnyway]

Yeah, I tend to agree that if someone goes out of their way to say one thing, it means that they are either doing another, or so likely to be doing another thing, that we probably shouldn't trust them.

Even on the level of customer service, Google is highly dubious in terms of click fraud claims and what not...

[fldz]

Cris, I'm glad guys like you are paying attention. We always underestimate future potential until we're in the thick of it. AI in war machines is barely even an issue and it's being developed without worrying anyone's pretty little heads about it.

[fldz]

And talk about the ULTIMATE Nuremberg Defense :)

[Dirtfire]

My logic is undeniable!

[SwampMidget]

I'm sorry. I can't do that, dave.

[p7tms]

Another problem a.i.would look at its environment and be confused. People created games that have an ultimate goal of kiling- and would a.i. think that is normal? It might get the impression that the ultimate goal of a human being (and itself) is to kill. Then, try to explain why humans have games that kill, why is it that some people get punished for killing, while others do not. And combine this issue with trying to explain what morals and ethics are and it is a difficult situation.

[p7tms]

There is another issue to look at. An a.i. would see humans as being flawed because of their emotional/pshchological issues: it might consider itself as perfect (which begins megalomania), and that anything that is not perfect needs to be fixed or destroyed if it cannot be repaired.

[SearchAnyway]

That's a fascinating point: even if its programmed to not be emotional, it would still have an ego. I have to stew on this...

[mistamustard]

There has been a lot of writing on the topic of how an ideal seed AI should be put together. Something you might note is that, like us, machines rely on hardware and therefore part of our thinking is influenced by the structure of our systems. Our emotions are geared toward survival and replication. It's possible to give an AI a similar drive such as "please the human masters."