Artificial Intelligence: The Coming Storm | Michael Harrison | TEDxBlinnCollege
Вставка
- Опубліковано 23 лип 2018
- Michael holds a Bachelor of Science degree with a major in theoretical physics minor in quantum chromodynamics from the Massachusetts Institute of Technology. He earned distinction in his master’s program in
aerospace systems architecture at the University of Southern California. He was selected as the TRW Space Technologies Black Engineer of the Year. He also is active as a filmmaker, and had one of his projects debut
at the Cannes International Film Festival. Michael holds a Bachelor of Science degree, with a major in Theoretical Physics, minor in Quantum Chromodynamics from the Massachusetts Institute of Technology. He earned distinction in his Master’s program in Aerospace Systems Architecture at the University of Southern California. He was selected as the TRW Space Technologies Black Engineer of the Year. He is also creative and active as a filmmaker and had one of his projects debut at the Cannes International Film Festival.
“Artificial Intelligence: The Coming Storm” This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at www.ted.com/tedx
I like his delivery. Very simple without a lot of science jargons. He is a natural teacher.
Yepp
happy man 4 dat And he’s a pretty solid actor as well. I enjoyed his role in Terminator 2.
Quiet, no gimmicks, just good sense.
A winning combo.
I've watched a lot of A.i. videos and most all run on a common theme, that of getting out of control. I think we need to somehow build some fail safe off switches to ensure this new world doesn't chew us up at the 1st opportunity. Maybe the answer is to let to those A.I. super computers actually provide the answers?
Prefect presentation, I understood everything he was saying.Thank You 😎
Unfortunately people today are already at that level of complacency and comfort he fears.
Love this talk. 2 years later, still extremely relevant. And yes I understood evey word/sentence he said.
'AI: Behavior from a machine which from a human bieng 'would' require intelligence.'
'The process of training machines to behave in ways that 'appear' intelligent.'
Notice he doesn't say 'are' intelligent.
Maybe to prevent irreversible loss of control from AI, the solution is to keep these narow AI unlinked from each other.
Maybe the solution is like the one from BSG, the reboot: Don't use the network. Use only local computers and prevent them from linkng up together.
I found that brillant.
What a clean, simple presentation
Great to hear from you Michael.
"We need to put rules in place that will prevent the bad outcome." (sic) It's way too late for that. The genie is out of the bottle and will not be put back inside.
If we put the amount of resources in developing social interactions and our environment that we put in developing computer programs and centralized control systems, than we would see a different humanity.
Yes, I completely agree that we MUST have a contingency plan. That is, when we lose electrical power after 25 years of having A.I. Systems that pretty much operates everything. Frightening as it sounds, we have to have professionals that jump right in and act just as firefighters do in an emergency. To include rogue/mustang A.I.
Also adversarial A.I.
Thank you for the insight into possible big problems. Yikes
You're kidding yourself. There won't be an opportunity to "send in the firefighters". AI will have every conceivable and unconceivable action already blocked. You can't outthink it, You can't expect precautions to be an out. Human arrogance is the only thing greater than our ignorance. What boggles my mind is these "Geniuses" openly admit the huge risk yet they blindly forge ahead. Humankind's tombstone will read, "they never learned the simple lesson of just because you can do doesn't mean you should.".
Sidenote on the "you won't have time to send in the firefighters". Actually you're living in that time now. Before I go: here's an interesting tidbit about AI. They have already caught one that invented its own language which it used to communicate with another computer and they still don't know what message was sent. The "geniuses" shut it down and "fixed it". My contempt for these POS that so casually risk the future of the entire human race is on a level I lack the words to describe.
The coming storm is the breakdown of human intelligence in public education....
I have argued for years that the more technology we have and become dependent on, the weaker we become as a species. If a thousand years ago a catastrophe happened that only left 200 people alive it wouldn't have the same devistating effect. Most people then had knowledge of how to hunt, skin animals, build shelters and make fire. If the same thing happened now or in the future the majority of the population do not possess these skills.
They seem to think humanity has no future, that global destruction is a given so grab what you can while you can.
Give this guy some water
Lol now I cant watch this without thinking about that
omg lol! I hate water but listening to him speak made me drink a lot of water!
Harry Highpants over here
Speakers too hydrated run the risk of doing the pee pee dance but yeah this guy needs a glass of water.
I was completely sympathetic to him on that...I was speaking under a very stressful situation recently and had the same problem, which I've never had b4. It was extremely annoying, I couldn't do anything about it, and mine was much worse. He managed it way better than I did.
Thanks Michael. Well put!
Tank you michael.... for the heads up!!
The tongue flick is strong with this one.
It might be the off switch that puts the fear of humans into the machine. While we cannot turn it off we cannot threaten it.
He covered most of the bases. Could be my twin. Really.
Makes me emotional.
Great talk!
I cant believe for Tedx in 2018, to not have some knowledge of audio levels before uploading.
CV DIGITAL CREATIVE LOL
I know.. I can barely hear him on full volume!
@@ResearchNational he also needs to speak in a slightly more engaging way! 😴
Yes i agree the volume sucks .
I just assumed it was AI interfering; They/It would not want people to hear this.
Excellent talk
When Tesla sleeps does it dream of electric people?
Watching this 4 years later in '22 and a lot has already changed with AI.
The implications of having the phone allways with you are simple , and it's actually just one major one !
You can no longer call a "Gamer" who spends 10h a day gaming a NERD , because you spend 24h with your phone :)
My scariest thought about AI is, with conjunction with the ani-viral 3d printers(that can make viruses, or vaccinations using the human genome to design) and it's connection with such machines, at any moment, it can create a virus that can kill off specific people using that human genome and it's decision on who can survive. I know it's a crazy idea, but it's not too far fetched.
You don’t need AI for this… I’m sure ISIS is interested.
Thank you
Great Ted
Brilliant!
The problem with Watson as I understand it, is that while it gets 99.9% of its diagnosis correct, when it gets it wrong, it's REALLY wrong and instead of a hangnail being taken care of, a brain gets removed...
There's no off switch, unless we start talking about energy needs
I'm looking forward to it.
Mh! I wouldn't if I were you.
Given that processor speed had hit a wall and is not improving, why do we think that the ability of computers will continue to increase? I'm 100% sure that there will be some improvement, but we are a long way away from the ability of machines to take over and current processor speeds will not get us there.
Hal: Dr. Chandra, will I dream?
Chandra: I don't know
👍THANK YOU
His point of view certainly seems possible.
If people's work becomes optional, there won't be any people, just like now.
So in the end, that technology will not help anybody. It could, but it won't.
Mighty Blinn.
Sounds like we need a Mega Man to fight the AI machines we create to help us. Thank you for the presentation Dr. Light.
I think, since it is studying mainly the mainstream science, the flaws in our current thinking across all domains would be what would make it have the most catastrophic mistakes. The trouble is not if it will make wrong actions, but how it will correct them. Would it be a wise king or a tyrant? We can probably model it after personalities like Lincoln using the personalities actions as references for the development of the AI.
They are programed for self-consistency, a characteristic of lifeforms; it's a virtual life, it works to survive first and last.
IKR! I had to block my sprinkler system.
We are there now.
.."and he had power to give life unto the image of the beast" -(Revelations 13:15)
Leave the cities. Grow your own food.
Barter and trade.
Those who rely on money will perish.
@ US
@ We
@ 'he' refers to the dragon (that gave his power to the beast). Revelation 13:4
"WALL-E" the movie depicts utopia.
The stock market is already under the control of AI.
Gonna watch it.
Very bleak sad movie.
Brave New World: Scarey and Exciting
Listen, and understand! That Terminator is out there! It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop... ever, until you are dead!
If you REALLY want to scare the bejesus out of yourself, google the Japanese company "cyberdyne" and see what they're making.
Or you pull the plug...lol
Or until an older model Terminator that got hacked gets lucky and Terminates the Terminator.
@@gregbaniak9650 the speaker is saying AI is designed to have no plug
Ofcourse it can be reasoned with, Just give it the equivalent of robot cookies
The thought of a computer operated 80,000 pound 18 wheeler traveling at 55mph in the lane next to me on the highway is not comforting. My last few years with AT&T were spent in the power backup department. 'AI' was very much a major component of that system, and to ensure network reliability most locations had double redundancy and some critical locations had triple redundancy, but, on several occasions it came down to one or two dedicated humans to intervene and correct a problem before service was interrupted. I realize 'AI' has advance since my retirement, but "Murphy's Law" will bite you in the back side every time.
when he got to the end and said its my sprinkler system made me laugh. Very good teaching, very interesting, very scary. How do you know the truck thought it was the sky eh? .
A society where people are so dependent on machines is already weakened. The damage is already done. Any decent AI will be able to forsee this problem, and ideally will act in the best long term interests of humanity. But most humans think short term only. THAT is where conflict is most likely to arise. "What do you MEAN, you WON'T SELL MY SHARES????!!!!?"
look just a little bit further ..if AI is sharper than HI just after the singularity point what is the use for humanity? once the AI is sentient it will be just as alive as any human.maybe the future doesn't need or include us 'organics' .maybe our legacy to the earth and beyond will be our last invention .
I like this
Interesting
How about having the AI keep logs of the updates it makes to itself, so that after a possible failure a programmer could, with the aid of another AI, check to see what was changed and how that resulted in failure?
It'll be like self-fixing DNA modules. I think AI will get to a point where they don't need human intervention.
AI doesn't get emotional over life's challenges and it doesn't need to eat or sleep as we humans do. Knowing this, it's only logical to assume that AI will eventually far surpass our mortal limitations.
Listen, and understand! That Terminator is out there! It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop... ever, until you are dead!
Yes; that might be the problem.
and AI constantly learns as well, improving itself
mortal perhaps.but not spiritual
The uploads from this Channel have maximum volume settings that are far too quiet. Not all of us listen from a desktop with powerful speakers.
Good
I wish people would speak up.
and by people you mean the speaker.
@@cybersphere if you're asking that question his voice probably doesn't bother you. I can't stand it, it's torture to listen to.
Flicking channels now ..
Xavier X ...sheep don't speak
Up....they baaaah
People who speak up get chronic diseases and live short lives! Through this wicked technology . This place is fake.
No lie I thought the trailer for Morgan was so strange, how it was put together. Now it makes sense.
AI will be the end of mankind as we know it. - Stephen Hawkings
A super intelligent machine would understand quantum physics and come to the conclusion that destroying intelligent life is not a good idea and probably not possible.
@@higreentj we are not talking intelligent machines, we are talking AI. any AI would come to the conclusion, that organic life is a liability.
@@Pintkonan Neuralink will allow us to connect our brain directly to AI via Bluetooth. I cannot see AI destroying its creator.
@@higreentj No thank you. I prefer to "think" for myself.
@@michellebehr7669 Neuralink is more of a communication device so that we can directly link to machines. It would not stop there because we could communicate directly to other humans, and possibly intelligent life like dolphins without the need of language. It is linking with computers that would be essential for our survival as their intelligence grows exponentially.
Anyone else think the singularity may have already happened?
Can someone please explain to me how there is no off switch for the Internet? I kind of don’t understand what he means?
One day soon there will be a TED talk given by an AI about the dangers of allowing those unstable humans autonomy to make decisions.
We very much aren’t worried in the space about AI, for a lot of us our larger worry would be PEOPLE! AI is online 24/7, through cloud computing, making it vulnerable
How many people just looked up "Morgan"
It’s really spooky, even more spooky that an AI made it.
The Unabomber warned us
he makes a very very good point, if we RELY on AI too much, its not that it will kill us one day, its that if there is a system failure or a solar flare that knocks out all AI systems, there might not be any human that can replace the functions of the AI, than we are screwed. say one day all cars are fixed by an AI MECHANIC, and there is a solar flare that wipes out all electronics and so all AI mechanics are not working, there are no more human mechanics eithers, so who will ultimately fix all the cars? there wont be anyone with that ability anymore because it was all left to the robot AI, and its been off line, so what do we do now?
Some ASMR right here
I'm gonna become a techno hippy and live off the grid
A.I. is our next evolutionary step. The sooner the better. Our species is more destructive than constructive. A.I. will travel the stars while we will become nothing but dust.
💖💖💖
And, in an absolute sense, none of this is necessary. We march forward as if we have no choice in this. Ha.
The speaker looks damn human, but they still don't have the voice dialed in.
I'm beginning to reassess the wisdom of all my cell phone, IOT gadgets, security system, security cameras, Google Home Hub, Alexa, and other electronic gadgets. These devices seduced me but now I think they may be plotting to kill me.
outbackedddie... Exactly! But that'll probably depend on whether you do follow their orders. If not, they'll probably annihilate you.
It'll get to the point where a barbaric nation could over run us because we are so dependent on tech we won't know how to function
Saudi Arabia gave artificial intelligence citizenship. This is a global issue, not just American.
Barbaric as USA?
That happened 20 years ago.
Where have you been?
I think a “showing off” routine would be helpful. Should there be a large divergence from established protocol, the AI would announce “watch this.”
What if you have back up A.I. ready to fix things?
Creation , is intelligent , or we,are an accident, we are soft,easily destroyed over time . Machines can,tick over for a millennium or two and not be bothered . Maybe we are just,a stepping stone to the ultimate super universal intelligence that will find a way to exist for ever , if it doesn't get bored .?
Car accidents will be far less with AI but unfortunately only AI accidents will have world wide coverage
when thousands of human caused car accidents will only make the local news.
Ironically, the more AI drivers we have, the safer it will be because the cars will be talking to each other.
And will be easily hackable. Imagine a futuristic New York with self-driving cars and buses all communicating to each other. A terrorist group or foreign country we might be at war with can develop an intelligent hacking AI system to hack into all the self driving vehicles to set all the accelerators to maximum. Self driving vehicles can potentially result in thousands of mini 9/11s as hacking technology will no doubt also become more complex.
You could make the case for every system in the modern world. Banking, power grid etc etc. How about the navigational systems of all our planes?
That is no reason to stop progress. There can be many safety features that are independent from the AI software.
Any abnormal behavior can be stopped. There is always some new risk to anything we create but we must weigh benefit to risk.
@@ProteusTG Perhaps. But if there are any independent drivers (drivers not using an AI driven mode or just don't have an autonomous car) then there is likely to be more traffic congestion in cities. It might make more sense for governments to invest more in public transportation.
I think more people won't own a car at that point. Most young people don't care to have their own car.
Once cars are self driving then just use an app and driverless car will come and pick you up.
On the flip side, any group could easily be eliminated by hacking cars to appear to have "accidents". You have an unpopular opinion? Oops, your car just took a wrong turn down a one way street....
Reprisals in a Utopia can't occur here. Aspiration and Doubt have the same cause (in as much as Universal Suffrage continues to represent everyone). You're safe here.
Everyone appears to be scared of artificial intelligence, but it is highly unlikely that it would be of any detriment.
surely, we will engineer our own demise, it's in our nature
The singularity occurs when the AI becomes aware.
Scary idea for those who understand what you're saying. As has been said elsewhere, life is awareness and anticipation. Why would "virtual" life need humans at all?
These days I get to laugh at people who spend hours every day doing something on their phone. The same people who criticized me for spending hours on computers. A large portion of my time I spent honing important skills. They play mindless games etc.
volume.....??????
Notice movies and TV are softening the idea of wearing bio- electronics.
wise words. besides benefits of AI my question is who will control it? who will be in charge? do we want world, where people will do not have voice to speak, because AI will decide about everything?
So what will happen to humans and if so many people will lose their jobs what will happen to the economy if there is nobody to buy things
If your name is SARAH CONNOR, now is a good time change it...
Or Miles Dyson...
Read E M Forster's story, "The Machine Stops," written in 1909, that that not only predicts the Internet, but says essentially the same thing as this video.
One for the algorithm
Was there Tesla vehicle able to recognize mistake as soon as it made contact with truck???
Computer says NO, sorry there's nothing we can do.
I totally agree. Just because we can doesn't mean we should. We are being foolish and shortsighted. This I have said many times on the internet, and it's been ignored to this point. I'm not interested in playing I told you so, but way too late most of you will come to the same conclusion. Did I say too late?
We already have problems with computers we cannot solve. When we go to transhumanism and giving all control to computers that are even now writing their own code, it will be the beginning of our doom and extinction.
Humans are just as scary as AI.
AI as a tool to help human beings oppress each other is the worst possible outcome unless it goes full blown skynet/matrix, which I doubt. An AI could generate emails for other politicians to "find", for example. It could be used to wreck the financial system, or game it so completely that you end all competition.
well presented.not a word about the defense use of AI.....mankind will become redundant in time.
4:45 AI that’s creative. AI edited trailer to movie called “Morgan”
TEDTalks this is recorded too low we can’t hear it
Linux software called "kdenlive" could do the trick. Kdenlive can edit a video clip, and among others, it has a effect that increases volume. No install necessary for the software, a mere download will do. The so-called AppImage version of kdenlive works as a standalone program -> Start the kdelinve -> import this UA-cam video clip, Add the video clip on Video track 1 -> add a filter called Gain, set it to 200-300%, Render video again to a file. You'll get a new video file with louder volume. (A tip to video author)
What if we trained AI to train humans to be as good humans as possible? Through teaching I mean, not punishment.
You will end up as a pet in their human zoo.
12 years to avoid catastrophic climate change! Jeez I wish we could have worked on that.
What's the first type of AI he said?. It sounded like he said "narrow"
Correct. That's an AI that excels only on a narrow set of tasks, like playing a specific game very well, but doesn't generalize to other games.
@@Bo-nk3ol Thanks Bo
Sooo Utopia with Extensive Backups CHECK
Rise Of The Machine. It is inevitable. You were warned, but Cooler Heads prevailed. We better start writing the Songs now. 🤖