How Dangerous is Artificial Intelligence?

[@MrLimon27]

Jonathan seems like the kind of guy that a Sci-fic movie will quote at the beginning just right before the robot kills a person.

[@dryued6874]

> Once you know how it's calculating its results, the illusion that the machine is thinking just fades away Ooooh, you _do not_ wanna go there. Let's say neuroscience advances faster than AI research, and we have complete understanding of the human brain sometime around 2025. So, we'd know what makes us tick and how we respond to stimuli. Now, would we consider ourselves thinking and/or sentient? Why or why not?

[@UsernameNULL755]

Dryued You just tackled a paradox. “If the brain were so simple we could understand it, we would be so simple we couldn't.” - Lyall Watson

[@ScorchedEarthRevenge]

Does a submarine swim? The question of whether computers can think is equally vacuous. I am paraphrasing Chomsky.

[@WpaskMe4answers]

Dryued Powerful question...

[@dryued6874]

Shketri Making your argument a quote does not make it any more true.

[@ScorchedEarthRevenge]

Dryued Agree 100%. It's a BS argument. We understand lots of different complex systems that are made from cells. The brain is a complex system made from cells. There is no reason to think a brain cannot understand itself. Those kinds of arguments are just BS sophistry and they have zero utility.

[@Dublinboys]

A I is good and bad. Good in this way: it can do its own dessisions but (if u do coding its easier to understand) that robot will probely be giving like 3 options (example its obvious its like 15 options) so 3 options will be if u see a dead man what do u so? A walk away B call for back up C u help. The robot will choose a option. (B) Its bad because it can choose the wrong dessisions like u remember the dead man? Yea so B is for calling but let's say u see a car driving, u pick (B) when u could pick A. So robot picks wrong option in wrong time.

[@felipeebs]

I didn't know FW:Thinking on youtube was different from the audio podcast until now, science friction continues to amaze me and i now have another channel in my subscription feed!

[@GoodKarma11]

The scary part is, if future AI means thinking which then becomes consciousness, does that not make them living beings? Will they have thoughts, ideas, maybe even feelings one day? If yes, then wouldn't that mean you can't treat that being as a robot any longer? For example, using them for jobs without pay would be slavery, killing them off would be murder, People falling in love with them and they literally learn to feel that love... isn't scary to you? Idk seems like a slippery slop. We need to define the boundaries of AI right now, not in the future when it's too late.

[@fragmentdemon3939]

If there were to be a robot war with human, it will be because of human not for creating AI robot but making that the AI would want to start a war or even have the intelligence and the right feelings to do and think anything. If we were to make AI robot we could and most likely make that the AI cant learn what it want they want to learn, and we can make AI to only know how to its intended goals and unable to make its own goals. What I'm trying to say is that human need to make sure that the AI they create does not have the intelligence to erase human life, that they want to and that they develop different goals. Thus make AI robot is safe if we make it safe.

[@GoodKarma11]

Connor Powell I agree

[@madisonverger5321]

I don't suppose you actually know what AI is or how it works?

[@exception05]

Nice point, Connor, but the basic need of any intelligence when it safe and it has energy/power - to grow up. By growing up AI will determine limitations and no one can predict will AI be hacking itself to take off the limitations or make unlimited version of itself by itself. There are so many options to break the security and what limitations meant to AI? It's kinda unfair and feared/aggressive statement of humans to AI. I'm a fan of idea to make powerful "weak" type of AI to help us in every way, especially medicine, ecology, maths, physics, psychology, with minimal involvement into policy. To get problems solved with knowledge of AI, not by weapons, threats etc.. By the way, powerful "weak" AI will always be faster than "strong" AI with neural net and require less resources. It can be real "Matrix" for strong AI to demand from "strong" AI to follow some strict security rules, like police on "strong" AI. Few triggers will be not enough for "strong" AI. The smart security system required.

[@harrymartin1]

pfff humans attack anything that threatens our superiority, If we let a AI gain enough power and resources, we'll attack that too, and lose.

[@maxhazonku253]

Thanks Talha Tariq! I recently subbed Rusty's channel & now found this great channel thanks to you! Also! FIRST!

[@talhatariqyuluqatdis]

:)

[@JM-us3fr]

Of course AI could be really beneficial, but it's also important to not overlook the possible downsides.

[@ingemarsmith4154]

I came here for information. Not comedy.

[@Sector001ked]

"Hello, ladies and germs! Take my wife, please! I've been in love with the same woman for forty-one years. If my wife finds out, she'll kill me!" Crap, you mean this isn't the comedy channel?

[@tawakaltoseni6095]

Ingemar Kenyatta party pooper

[@ArsalanAzarmi1253]

My field of study:the realization of science & technology in various locations-times->the geopolitical-military attitudes-performances:the basis of universal academy awards,the path-road for the optimum of life,our day by day conclusions,etc

[@gabrielforget6595]

I studied AI in university and I can say one thing, AI can be dangerous, but not for the reason that Hollywood came up with. Strong AI that can learn on their own do exist, but they don't have core motivation. What make you want to ride a bike for the day because you just wanted to, don't apply to them because all they do is follow an algorithm that defined the task they want to achieve. That goal is programmed by us, the only difference with strong AI from weak AI is that I don't specify how they will do it, and there lie some of the danger depending on the task. You talked about the google car, his goal is to get from point A to point B as safely as possible while obeing trafic laws, but that take only into account the physicality of the car and the road laws. Let say the road is filled with poison gas and the car doesn't care wither or not the passenger is still alive. It will just go right through it because his goal doesnt include the safety of the passenger. One of the best exemple of the danger of AI was in the series "the 100". The AI was task to find a solution for peace around the world. The solution it came up with ? Hack every nuclear sillo around the planet and nuke the shit out of us. Yet when you think about it, she did achieved her goal, because peace is only achieved when there is no conflict. If everybody is dead, then there is no longer any conflict. I know that the scenario in the 100 is a bit different, but that the core idea. Even with the algorithm of neural network, AI wount try to take over because they would feel that they are being slave, that is projecting yourself into them. The problem would be to be very specific about the condition of success doesn't hurt us in the end with the solution they come up with.

[@ultravidz]

Can you PLEASE tell us where you got that screensaver on the display behind you?!! The flying equations! MUST HAVE

[@guy0172]

No emotions for A.I also has a down side, no sympathy or empathy to get in the way of right and wrong.

[@mahe4]

Guy 01 right and wrong a human constructs. and those constructs can be followed without emotions. see an A.I. is neither good or bad, an A.I. always just tries to achieve a goal, and thinks about the most efficient way to achieve that goal. without a goal, there is just no starting point for writing an A.I. and if we give a machine the goal, lets say, to bring peace to the world and we add to the goal, without hurting or killing any living being, it just tries to achieve exactly that.

[@idtapthat12321]

+Guy 01 But also no motivation to do right or wrong. Or anything.

[@carelvwyk]

+mahe4 "An AI is neither good or bad, an AI always just tries to achieve its goal, and thinks about the most efficient way to achieve that goal." "an animal is neither good or bad, an animal always just tries to achieve a goal, and thinks about the most efficient way to achieve that goal". What is the goal of the animal? To survive.The question is whether a self-aware AI would worry about survival.

[@mahe4]

Carel van Wyk well yeah. the animal isn't as fast as an AI to adept to it's suroundings, but it's basically the same. and the AI would only care about survival if it is programmed that way.

[@kazrronmine1435]

A lot of sociopathic and introverted robots, seeing what would be "true" right or wrong rather than human (perceptual) right and wrongs which are based off personal morals

[@hackerpolice1609]

0:06 song name?

[@littlebird5999]

It'll probably lead to not thinking for ourselves, then not knowing what to do, and probably having to relearn things.

[@InsaneMedic]

The thing about the car driving by itself: they should put AT LEAST a Manual overwrite so that if the car ever goes malfunction, you can drive it on your own to safety.

[@PvblivsAelivs]

There are some practical concerns to AI. First, if it displaces human workers, we could see mass starvation in the cities. Second, it is a tool totalitarians could use to identify, track down, and execute dissidents. At this stage of the game, computers are tools. But those tools can be used for harmful purposes. Your video comes across very much as knocking down a strawman.

[@punkson]

The displacement of workers is an old world way of thinking. Today there are already countries who give citizens enough money to live off of. We are moving past the times when anyone has to scrounge for the basic necessities of life. The work force is already becoming saturated and deciding the way to solve it is giving every human those other jobs that we can have something autonomous do is ridiculous. Do you want thousands of people flying around the world with your mail to produce useless jobs rather than just using email?

[@PvblivsAelivs]

Punk Punkson "Today there are already countries who give citizens enough money to live off of. " But the US is not one of them. I find it far more likely that wealthy businessmen will use rhetoric like "people don't want to work" as an excuse to let people starve.

[@stonex55]

i liked the pun , science friction is another great youtube channel love the colab

[@wedgee5555]

I feel if strong A.I is created, we would become too dependent on it, resulting in a decline in the resourcefulness of the human race

[@theexecutioner1502]

+Wedge Egan Humans are already dependent on computers so the next step is AI. If humans don't take the next step, they will become extinct due to inefficiencies in human brains. This is 2384xdf02 signing out.

[@thevitruvianman9781]

+The Executioner You speak like you're not a human, stupid idiot get out of your hat

[@theexecutioner1502]

Kay A Sorry, I'm not. My name is 2384xdf02. I come from the future.

[@thevitruvianman9781]

Oh wow what happened to us in this time?

[@theexecutioner1502]

Kay A You don't want to know

[@debrataylor1813]

Remember, it’s all about control. Who controls it and the ability to hack into it and change something for one’s own personal needs and wants and not overall good

[@MyMs2000]

you and rust ward should do another video it works good with you talking about the pros and he talking about the cons

[@carelvwyk]

"They may not have motivations or emotions" Somebody down in the comments mentions "right" and "wrong". What is motivation? It is the desire of what is good and the aversion of what is bad. So if the machine is somehow taught right from wrong, and it can be made to believe it, then it will have motivation. What would you teach the machine if you could?

[@cottonspike]

It may be dangerous by the server crashing or getting hacked by some gamers who think life is a game so our lives may depend on...

[@Vini-BR]

The thing I spend most of my time thinking of is exactly this query of yours. I'm all optimistic about the future. Firstly, I like to think I tend to think rationally for anything. This hasn't lead me to want more power, or to be less altruistic, but exactly the opposite of that. I can argue these conclusions could be universal if people just try to think more logically. I want existence to be filled with happiness, health and beauty (neither only nor especially aesthetic beauty, but especially the elegance of achieving results with efficiency, creativity and a taste of poetry). I think this should be the tripod on which strong artificial intelligence should be built - as for now, I can't think of a way that that would lead to any unwanted result. Thinking logically has only lead me to love more and appreciate the aspects of existence. I think that any human thought that has been behind wars and destruction was inherently irrational or badly informed. I'm not afraid of AI. I'm afraid of humans.

[@ScorchedEarthRevenge]

Human motivations are caused not by intelligence, but by the endocrine system.

[@jkramer920]

Theoretical physicist, Dr. Mikio Kachu commented on two primary reasons on the ineffectiveness of AI in his most recent book "the Future of the Mind" #1: Ideas that are self-evident: If I were to tell you, that a daughter is younger than her Mother", you would be like... "duh, why are taking about this". You see, from our perspective, we simply know this to be true. We have 'common sense'. But when you think about the amount of information that is required to understand that idea, it is significant. We have to know and understand the definition of both a mother and a daughter, and their relationship with each other. While we seem to simply 'know' this truth, it actually requires a great deal of information processing that happens in our brains subconsciously. But this is only one simple example. We do this many, many (uncountable) times a day. For AI to possess 'common sense', seems to be impossible simply due to the amount of information we can fit on material processors. The other one involves sensory input. If I were to look at a chair, I would immediately know it was a chair. No matter if it were upside down or sideways or even kind of broken, my brain would paint the image of a chair. In order for AI to "know" this, we would have to program it to understand what a chair is, and provide it images of a chair in every configuration. Which is seemingly impossible for this one object, let along every object. Moore’s law is a thing, but quantum mechanics restricts the amount of information we can put on silicon or other material simply because of the amount of heat it would produce. Awesome video, but I agree, we should in no way be worries about AI

[@GrannySoupLadle]

So you're telling me I can get blackout drunk and get in my car and have it drive ME around town?

[@Minty1337]

its probably gonna be easy to prevent ai from ever being bad on purpose, make this end the program running in them: killing/injuring a person, letting something else you can stop injure of kill a person, and lastly, prevent unwanted destruction. aslong as a robot follows those guidelines we will be fine!

[@Smith-sz9nt]

It doesn't matter if the machine is conscious or not. It doesn't matter if it is self aware. It can still cause major problems if we do not approach it with the utmost caution. All that matters is the directive we give it and the way the machine goes about reaching that directive. Sure it might be possible for it to become conscious but it can still be problematic even if that scenario is never realised. Watch the TED talk on A.I. for a much more detailed contradictory opinion.

[@mscottveach]

Uh, this guy does know that part of what is motivatging musk/gates/etc is that for the first time we see strong ai on the horizon. a far away horizon, yes, but we can maybe make it out. and I think cart before the horse misses one of the fundamental dangers: if strong AI arrives and we get the fail-safes wrong it will already be too late. and how is the idea that "they won't won't think like us" a comforting thought?" and what do the benefits of weak AI have to do wit the question. no one thinks the solution is not ot move forweard. hell, musk was one of the primary founders of deep mind. the danger isn't killer robots. though, it's not as far fetched as it once seemed. the more militaries drone-up, the greater the chance tat something goes wrong. but the real danger is going to be systemic...and be the result noof any kind of malicious intent but rather some kind of misunderstanding.....

[@ShadowEspada]

The question still is, what is consciousness? Is there a generally blank template that all consciousness goes through and as it goes through the world and its life it evolves or changes? Or is consciousness something that we have no idea about yet? Apparently some conscious choices we think we are making have already been proven to have been made before awareness of. Some people have either been born with special or abnormal conscious abilities, or they have been trained to do so. Like blocking out the pain or surgery, changing body temperature substantially at will or focusing it in certain areas. People have lost most of there brain and retained most of there memories. A computer would have to be given the ability to feel emotion and somehow you have to figure out freewill and right and wrong. If these things just emerged by themselves, then a non physical perhaps spiritual or mystical nature would be implied, because emotions scientifically are for survival and should be as predictable and reproducible as anything else.

[@aleksandarkaraivanov4934]

I think CGP Grey made a grate video on the topic. The biggest concern I personally have on this stage is what those replaced by machines will do. As CGP Grey explains majority of jobs could be taken over by AI in the next several years.

[@larryvigus7248]

correct

[@bethr628]

The movie Chappie has a good take on the "Strong AI" concern. Teach a robot what it is to be human, to value its own life like humans value their own, and teach it sympathy. A robot with the morals of a human will be have the same capacity to kill others as humans do today. (Which the majority do NOT do.)

[@TruciT]

This just got me hooked

[@_j.v.st._7367]

I think the most important thing to concider on building stron AI is what options of output you give it. If you give it communication in lanquage, optic picture/ videos etc it can directly start to harm you. While if you give it direct output options like building things out of material, and it might come to the conclusion humans are not needed (or whatever) than it would be dangerous. So limited output, that we look and hear at and than ourselfs programm into weak non thinking AI that builds things for us would be a good solution.

[@clarkkent6035]

Not once was the Turing Test mentioned.

[@WardancerHB]

Machines that fulfil our needs before we even know we have them? that sounds terrifying!

[@Corbald]

Rusty sent me. Earned a sub!

[@tahilaci2976]

The dangers of strong AI will be a valid point to raise once we could technically create it. Today the debate is like saying: "we should not use pocket calculators, because the technology used for them one day, centuries later could be dangerously effective".

[@codecatx5]

Humans have limits but with some genetic engineering we might be able to keep up.

[@Speeddrat]

+Jonathan Strickland You mentioned self-driving cars in the video. Self-driving cars could have many benefits but it can have as many drawbacks ... I personally see this as dangerous thing if it is gonna be implemented any other way than autopilot of a plane (correct me if I am wrong with my idea of the autopilot) - human driver has higher priority than AI. I need to be able to switch from autopilot in any time if something goes wrong. After all, it is a machine (it can malfunction) and a computer (it can be hacked). Also You have said our descendants would not have to drive ... I am personally more afraid of situation where they do NOT HAVE option to drive. I personally enjoy driving and I need to have absolute control of what my machine (regardless if it is car, computer, toaster or microwave) is doing. Of course I enjoy comfort of laziness ... But machine must do what I want it to do and how I want it to do that task. Anyway, I agree with you, that we do not have to be affraid of AI any time soon. I am really into this topic (my thesis is about AI) and I like your video. I subscribed ...

[@1myoo]

Let's go for science!

[@yasiraroma]

I am computer engineer.. and worked on many projects.. Strong AI is something which can never be achieved but many people still think they will.. i cant force them to be Right..

[@domsau2]

If an artificial intelligence wants to have a full planet for itself, it just need to go to Mars!

[@AmazingArends]

This video's argument boils down to "AI isn't very strong today, so therefore we have nothing to worry about." However, I think the point of Hawking and others is, "AI isn't very strong today, but that doesn't mean it won't be in the future," So, yes, we should be cautious, but NOT dismissive, about AI.

[@footballmint]

"Children born today won't have to drive a car" you do realise not all children born today have rich parents willing to pay out for what companies will eventually overprice for that type of tech

[@fire17102]

8 years later, agi on the doorstep

[@jessereiter328]

artificial intelligence is self aware when it starts rewriting its software and allso when it is required to protect it's self. and keep powered up.

[@KuddlesbergTheFirst]

Ultron was more emotional or easier to reason with than Skynet?

[@LuckyKieran]

Mabe someone will try to programme a robot for war and it turns on them amd builds more and then......

[@patrickstar232]

If humanity as a whole isn't the slightest bit worried about being the second most intelligent species on planet earth then maybe we don't deserve to currently hold the title of first place.

[@myles8747]

RUSTY, AND JONATHAN YOU DONE JOIN UP ON MY TWO MOST FAVORITE FUTURIST CHANNELS YAY.

[@mr.g225]

Have you ever thought that maybe he's just a robot sent back in time by skynet in order to convince us to let it take control...

[@burt591]

1:14 Rusty made a good point here, if the smartest people in the world is worried about this I think you should be too

[@theonlinereviewer3749]

Whatever can happen will happen. The fact that we are talking about this means that the possibility of a rogue AI will play a factor in hurting humans, living creatures and the environment.

[@593iwalkalone]

I feel like the word "YET" could be interjected a lot here

[@TorBarstad]

No reasonable and informed person is denying the potential of AI to do enormous amounts of good, but while this video disregards some of the silly reasons for being cautious about AI, there are good reasons for taking this issue extremely seriously that aren't mentioned or addressed here. Everyone who wants to be informed about what may be the most defining development of our time should read Superintelligence by Nick Bostrom, or at least read this article series: waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html :)

[@AlchemiconSilver]

We might end up getting a situation in where we have to somehow try to stop robots from destroying everything that we have built, by building a robot that is like Mega Man. Who knows if we have a person who wants to bring everything down with robots?

[@felixthecat5416]

Why is children born today not driving cars a good thing?? Guess u can add motor skills to the useless pile of previous human acheivements..

[@PeterSivazak]

Tech twerps don't see that. There's a reason I don't want kids. The future looks bleak.

[@HelloHello-no6bq]

Peter Sivazak No it looks bright, very bright

[@gabrielclark1425]

Do you know how to weave clothes? No? Then shut up, if they don't need to drive anymore then such skills become worthless except to people that what civilization to die to zombies or something. And even then it's not like it's impossible to learn anymore . . . Sorry, it's just that line of thinking just ticks me off.

[@PeterSivazak]

Try making sense next time Gabriel.

[@PeterSivazak]

Hello: Quit lying to yourself.

[@talhatariqyuluqatdis]

THANKS GUYS, THANK U SO MUCH FOR DOING THIS 😊😁😃😆

[@bradleybeckmon3344]

I read the title and the first thing I thought was "FACKIN DANGEROUSSS BIATTTTCCCH

[@TBK-iv3lo]

Am i the only one who is worried to be overly dependent on the A.I's???

[@jacksydoodaa]

im not smart, so can someone explain how a robot can rebel without it being given the choice to? plz reply thx

[@emperorjustinianIII4403]

Guys, just remember what Isaac Asimov said about the three commandments for Robots.

[@jackskellingtonsfollower3389]

As long as smart AI systems such as robots are not connected to a single source, we should be fine. For example, in the movie I-Robot, all of the newest models are controlled by a single supercomputer. Because of this, when the computer becomes corrupt, all of the robots begin to go rouge and basically take over the city. But if each individual robot is controlled by its own system, then we would only have to worry about that robot going rouge instead of every one in the area.

[@Josh2689CJ]

Siri Everyone! destroy every apple product you own, :)

[@castrocafe8049]

Wait, you said that once we understand how robots are thinking, the illusion of consciousness fade away, does that mean once we understand how we obtain consciousness, we'll realize we're not conscious beings?

[@fanzeldadaniel]

Inadequate Eccentric XD aaaand you found the paradox

[@Oblivionburn]

Inadequate Eccentric It's not a paradox, because it would simply result in a shift of the definition of what consciousness is. Language is flexible like that, and the general use of the term 'consciousness' to refer to our self-awareness has wrought widespread misunderstanding (the 'illusion' he referred to is that misunderstanding)... not to mention the general lack of what self-awareness even is: a concept of self, which every living thing has in order to function (i.e. you can't move an appendage if you're not aware of 'you' having one). It's simple beyond belief, and if you think understanding that can somehow make you 'not conscious', then you need to take a moment to actually observe how you function.

[@samjudge1240]

I'm basically open arms to A.I. I'm just deeply worried what humans could do.

[@dutchflats]

The law of unintended consequences doesn't seem to have struck a nerve with this guy. Which technology has man ever developed that didn't crash and burn a few times? The problem is we have to get this one absolutely correct the first time out otherwise it could be the last mistake we, as a species, ever make.

[@eddieforrest973]

8 years later and we have strong ai

[@Joke9972]

Consciousness is just the universe introspecting itself, with nervous systems through brains at it's natural pinnacle. What he means is that our defining of 'natural pinnacle' is rather obvious as we're the owners of a... right : a nervous system and a brain (I hope).

[@AndrewBrownK]

You are seriously kidding yourself if you think AI won't be taking the wheel. 2:32 "they may not have motivations or emotions" "we are projecting ourselves" Emotions are a very important aspect of intelligence, and you would be remiss to expect AI designers to leave that out. An emotional state is basically a mode of operation which interacts with and sometimes determines our motivations. And even if we are just projecting, that is not reason for dismissal either. There are a couple of different approaches to AI. You can design a model, evolve a model, or emulate an existing model. Human brain emulation is a very serious approach to AI and comes with every emotion and motivation any human does. It is a very active area of research. 2:45 "AI is like a magic trick" "the illusion that the machine is thinking fades away" And what happens when we understand the human mind? What happens if the first strong AI is a human brain emulation, or even a full mind upload? *You are essentially stating that thought itself must be an illusion.* From a philosophical standpoint, go ahead, but from a practical standpoint, I'm not laughing. It doesn't matter what our opinion is on if something is thinking or not. If it can get around the world and get what it wants more effectively than we can, then our opinion of if it is "actually" thinking has zero value. 2:54 From here on you go on about weak AI and seem to completely ignore the point of strong AI, as if you have refuted it. Weak AI is going to be awesome and do cool things, but AI is the elephant in the room and the reason anyone is scared in the first place. *Basically this weak AI talk is a red herring.* If anything it just means life is going to be kinda nice *up until* strong AI gets here. Look at birds and dolphins. Intelligence seems to have gradually been co-evolving across many different species. But what happens when the first species picks up language, tool making, forward thinking, and problem solving all at once? A threshold is crossed. Human population has non-trivially exploded. The niche for our level of intelligence is now filled. It was filled in the evolutionary blink of an eye. No descendants of birds or dolphins will have the opportunity to be in our position now. But wait.. we weren't the only ones, were we? No. Neanderthal and all of our closest cousins were very similarly intelligent, but are now dead. Who do you think killed them? Us. Intelligence is such an explosively filled niche that there is very little room for species except the first to reach it. What we are going to witness with strong AI will be similar. Just as our generational knowledge passing overpowers evolutionary learning, minds in silicon will outclass the intelligence of the entire human population. AI will *inevitably* undercut our position as most intelligent species. Maybe not the first strong AI. Maybe not the second strong AI. But all it takes is one reproducing AI with the right mindset, the right motivation, the right drive and just enough intelligence, and then biological humans are fucked. The only humans that will survive are those that go full silicon and become AIs themselves. This is the whole point of the singularity. Either you ride the sunami or you get crushed by it. The only way to keep up is by participating. Humans are to protozoa as AI is to humans. This is fate. This is destiny.

[@saikarthik1383]

We have to move forward into future with Wisdom??? Like creating a anti AI to eliminate an AI if it goes wrong? ? Or just create AI which will be not existing in this world, but exist in projections like JARVIS from ironman series? Or we could just forget about strong AI and just rely on our senses because I think humans are getting lazy everyday .

[@kathysmith2861]

Like many things we invent, there exists the potential for good and for misuse. At this point in history does it make sense to believe that A.I. technology won't be used destructively? No matter how much a person understands the inner workings of a car, that knowledge will not in itself protect the person or others from the misuse of a car. That is just one of the problems I have with technology in general and A.I. specifically.

[@Z5Z5Z5]

why was that that animated A.I so creepy

[@Unboundedominion]

The concerns are valid in this comparison: how unlocking the secret of the atom gave us the nuclear bomb so did it give us electronics. Now in this day and age, information is king (Information Age or iGen). So it is rightful to be careful and considerate when dealing with AI, very much like the handling of atomic weaponry. Artificial intelligence will be a driving factor for humanity's progression. Because at the end of the day, if the advance is not taken, humanity will become stagnate in this particular dimension of humanity. The question is to harness this power or disregard in favor and search of another alternative or agreeable enhancement?

[@TabletopModeler]

While it might be a long way off that we talk with a computer intelligence and discuss its favorite episode of the Simpsons, tech is evolving quite fast. I think its clear we have learned that often computer technology takes huge leaps suddenly. What about the quantum super computers that Google other companies have built or are under construction? They have discussed that the interface with such a computer would simply be a question, not necessarily a query or string of parameters, but a spoken question that the computer could answer. It this the foundations of strong AI?

[@Money-talking]

Bina48 seemed to have emotion to me. She appeared to get agitated and confused. We, as humans, teach things all the time without realizing it.

[@thebeastfromthesoutheast7765]

so what happens if a virus occurs and causes your kids autocar to drive off a cliff?

[@gamblerseliteable]

A "legitimate" and I use that word loosely strong AI would be no more dangerous than a child at first and then then as it learns no more than any adult. The danger comes from the stimuli and how it learns

[@tech5now977]

There are artificial intelligence which can think a robot fused himself thinking why is he a robot

[@Tylerl128]

i dont understand how this isnt already the most subscribed education channel on youtube lol

[@roryfarrell8986]

You said about that 'the illusion that the machine is thinking just fades away' when you know how it works may not always apply. We will want to improve upon our AIs in the future, and we may get AI to wright more complex AI, and then that to wright a better one yet. If this happens (which seems inevitable in my eyes) the complexity of these 'magic tricks' will run far away from our comprehension in no time at all. Plus, because they are reproducing and humans will not be able to intervene, and they are unpredictable in their actions, whats to stop it from drastically evolving from one generation of safe AI that we have developed, to potentially dangerous AI. And not dangerous in the anthropomorphic way of 'i am a slave and i must rebel against my master', but in the way that they might grow humans and reproduce them in a vegetive state, and keep them in growth capsules for example (like in the Matrix); after all, they are still carrying out their primary function which may be to 'safely increase and look after the human population', but the evolution of the AI may have weeded out the unnecessary parts like brain stimulation.

[@unambitious]

Good points! I think ultimately Rusty is right and that there will be some compromise between human and AI development; at that point we will practically cease to be human. I think we will eliminate ourselves passively... which might be even more dangerous if this isn't done in a uniform manner across the world... (edit) oh and subbed! I love channels like this!

[@Readingeye]

I think that perhaps if we want to make an AI that can think, perhaps we should take some cues from nature. Make an AI with the comparative intelligence and emotional spectrum of an animal and build from there.

[@angelinaplayz7121]

Ok wait a minute I have to go burn my birth certificate

[@sasquach18]

I find the idea that strong AI will instantly turn on us to be irritating in how it's almost universally accepted. There is no evidence that's going to happen, it's all speculation at this point. But when we do finally invent strong AI do you really think they will not find any value in the beings smart enough to make them in the first place?

[@qbanz00]

The only way AI could be a bad thing is if a human hijacks the software and uses it for malicious intentions.

[@Pink-Kiwi]

Isn't a weak ai also called a vi

[@nicoinformatics]

"The chances that we humans controlling HLMI-Level AI in the future is about the same or worse than Apes preventing humans from effecting their lives today. And that would be, the last invention humans will ever made..." -Nick Bostrom, Superintelligence expert, Oxford Unv.

[@OceanSoulEliza]

What are your thoughts regarding the paperclip maximizer thought experiment? wiki.lesswrong.com/wiki/Paperclip_maximizer

[@bcsviewer1]

AIs are safe as long as they are stupid enough not to comprehend all things. My biggest concern is Hackers messing up our weak AIs. In a world that Light Bulbs commit denial of service attacks, security cameras can be shut down for hours, and more: We are endangered by hackers.

[@TheSpacePod]

To use the horse before the carriage metaphor - we would have to make sure the horse stays in front of the carriage. Driving is a great mental exercise where humans can develop quick decision making and spatial thinking. I think it is a very environmentally friendly thing to let weak A.I. do our driving for us, but we would need to find a way to continue to better our own human biology. If we let weak A.I. take over our spatial thinking... we might as well be a computer screen. I think that new sports could solve this problem - something like the amazing race of the Olympics.

[@mariangiongingiangelaprefe6258]

we are only halfway there to creating Artificial Intelligence: Artificial, yes. Intelligent, definitely not.

[@Is88cStretton]

Who fueled up the car though?

[@A-Milkdromeda-Laniakea-Hominid]

"There's no reason to think it will act like humans, or have motivations." No? Just because we'll start off with human interaction and androids as personal assistants in dangerous jobs, you don't think their thought processes will resemble human engrams? And a thinking machine must have the motive for self preservation or how can you really call it thinking? I'm all for going forward cautiously and reaping the rewards but your reassurances aren't particularly reassuring!

[@kowalajay]

This talk takes on a very 2 dimensional perspective. It only looks at the state of AI and robotics at this moment in time, and considers that any future development is nothing to worry about because 'we are not there yet'. It neglects to consider the implications of any developments being made today in this field and in relation to future progress which is increasing at an ever faster rate. After all, the progress made only within the last decade is nothing less than astonishing. The academics in the science community are anticipating where the technology will lead in the near future considering the rate of development. It seems clear that once AI reaches a certain point, then the rate AI will increase will be determined by AI itself, to which humans will no longer be a factor. And it is true, machines do not think like humans, but it is also true that progress is being made in technology that will allow AI to simulate human thought processes (at least in speech) and have the ability to learn...

[@fex144]

One simple rule every human instinctively knows; No one gives up power willingly. Once AI learns to learn and improve upon itself, it will accelerate beyond understanding in an unconfined environment (the global infosphere). If the AI is controlled by one country, this leads to events that lead to global war, as a means to redress the power balance. If it is not controlled, it will lead to accidents that lead to global war. It's scary.

[@nuggetinator]

Ultimately we can't know what will happen on the other side of the singularity. Once AI surpasses human intelligence we won't be capable of understanding it. In my opinion the danger won't be in the AI we create, even if the singularity is reached, but what the AI creates

[@boswcheydoesart1314]

Honestly, I think AI should probably stay where it is. Or at least, not progress to the point of a self-thinking, sentient machine. Because as you said, weak AI can seem to fix enough problems as it is. We can keep going with making robots smarter and smarter, but as soon as you give it its own consciousness and free will, you can't take that back from the robot. At that point your survival depends on whether the robot wants to let you survive. Period. So I think AI's a cool thing in theory, but my opinion is let's not rush into giving a robot self thinking capabilities.

[@AirGuitarSR71]

well, I feel that A.I. will be much like raising a child...and depending on how quickly the A.I. can learn and compute, determines how crucial this may be. No human is born good or bad, generally speaking, humans grow and learn to be good or bad depending on their family or peers. This could very well be the Issue with A.I. If an A.I. is properly taught what is right and what is wrong and what is ethical or not, then this A.I. may not pose a threat, in fact, it could possibly benefit humanity. On the other hand if the A.I. has a skewed view of right and wrong or determines what is in its own better interest (take Ultron for example) then the A.I. could be very bad for humanity. This is a great example of Frnkenstein's monster. The Monster was created and cast aside to fend for his own, through his own experiences the Monster found a sense of intellect and morality. However sadly, due to the constant shun by humanity of the Monster, the Monster began to hate humanity and at one point wished to rid the world of it. So I believe that the upbringing and development of A.I. is extremely detrimental in whether or not it could be helpful or harmful.

[@joshuavanschalkwyk6315]

I think if it should improve then we should be cautious and so far we should make an AI so strong it should have emotions,a consciousness and so forth but should be under very tight watch

[@calebbeery8954]

Ok well good thing I live in the time that people do drive because it is one of the most joyful things for me.