Visit to get our entire library of TED Talks, transcripts, translations, personalized Talk recommendations and more.
The danger of artificial intelligence isn't that it's going to rebel against us, but that it's going to do exactly what we ask it to do, says AI researcher Janelle Shane. Sharing the weird, sometimes alarming antics of AI algorithms as they try to solve human problems — like creating new ice cream flavors or recognizing cars on the road — Shane shows why AI doesn't yet measure up to real brains.
The TED Talks channel features the best talks and performances from the TED Conference, where the world's leading thinkers and doers give the talk of their lives in 18 minutes (or less). Look for talks on Technology, Entertainment and Design — plus science, business, global issues, the arts and more. You're welcome to link to or embed these videos, forward them to others and share these ideas with people you know. For more information on using TED for commercial purposes (e.g. employee learning, in a film or online course), submit a Media Request here:
Follow TED on Twitter:
Like TED on Facebook:
Subscribe to our channel:
Read that as “The danger of Weird Al”
Why has not 1 person realised “WE TURNED AI ON YEARS AGO” IT WAS ALIGNED!! AND PUT US IN THIS SIMULATION ON A QUANTUM COMPUTER .. it gave us what we asked.. it stopped us destroying ourselves…and gave us an environment where we can be free without destroying the planet.. as the AI GETS SMARTER the simulation gets better and we get better technology… That’s why it’s not a future world simulation.. it wasn’t strong enough.. but as it gets better our simulation gets better.. the big bang was the simulation turning on.. at a quantum level the world IS Q BITS ffs how can you ignore.. …if the world IS made of qbits how are we NOT on a quantum computer??
@Lee Parker : You generated this on a computer..
Same lol
That man is a menace
Me too!!!!!!
the AI seems like my toddler nephew that repeats anything i think is funny….much to my sisters dismay
It means A. I. is far from perfection
So…we’re going to innocently design our own destruction…that seems appropriate.
What I learned today
– the biggest fear is not AI malfunctioning on it’s own, it’s our own programming and design of the AI program
@Francois Roewer-Despres They can’t now, but they will be able to soon. You’re an imbecile if you think that AI can be controlled, or regulated without serious laws to what people can build right now, or how they’re built.
“It’s not happening now, so it won’t happen ever.”
*The problem with all human-designed systems is the limitation of humans to foresee things that the system can do, that aren’t planned. See the Verrazano-Narrows bridge as one of the original examples. But also spam, identity theft, social media, nuclear energy, and plenty of others.*
@Swizzy o7o7 You are the only imbecile here. You have at best a layman’s understanding of AI, fueled by so-called prophetic and apocalyptic warnings from people like Elon Musk, without an even basic knowledge of neural networks or deep learning. A.I. is not even close to being able to rewrite its own code in the present day. And if you watched the same ted talk I did, you would know that she is talking about AI in the here and now, not in a future with superintelligent AI. So the original commenter is correct in saying that most problems in AI come from the humans, not from the algorithms themselves.
Swizzy o7o7 I never said it can’t be done. Only that it isn’t being done now. There are some researchers developing AIs that help humans write small code snippets, but they aren’t rewriting their own code.
Will AIs rewrite their own code one day? Probably. But it won’t be done by well-intentioned people who know better, but rather by criminal/underground operations that think they can control such a beast. Unfortunately, regulation is never going to help you avoid that. Even if you made all AI illegal, the technology will just move underground.
In the meantime, AIs today would not benefit from being able to rewrite their own code, because each code modification entails days to weeks or retraining, making it an extremely slow process that a human can terminate whenever. In addition, the human can limit the type of code the AI is allowed to write/rewrite.
Basically, neural networks, which is the basis of all the AI hype of today, are not the right technology for developing super intelligence by means of self-rewrite unless you have a bad actor who is really, really trying to make it happen over many years, and even then it’s not clear that’s it’s even possible. Super intelligence will end up using a different technology that hasn’t been invented yet.
Another 15 to 20 years left before we’re no longer the most intelligent thing on Earth.
Nonsense , we have you.
@Marcus Rosales There is a study where a few hundred cutting edge AI researchers were asked to predict when general AI would come about. The average answer was a 50% chance happening in 2060, so more like 40 years.
More like 40-50. But that’s still nothing even in human time scales.
@Analytical Thinking Those scientists probably know a whole lot more than you… Moreover, experts in AI, outside of academia, are concentrated here in the US. Your scenarios are very, very unlikely… An appeal to ignorance.
I’m getting my PhD right now and talk to people working with AI; all of whom think a singularity in 10-20 years is media hype.
You posted a very defensive comment, which is not very ANALYTICAL! I wasn’t saying it won’t happen, but I’m going to trust real scientists over pop science, which is more than likely where you learned everything you know. Sorry if that sounds rude, but that’s what I think.
@Lightning445 That sounds about right. It’ll be an exponential increase, like all research, so it’s hard to say when for sure.
I just think many of these people predictioning GAI on YouTube have no experience in coding whatsoever, and saying 10 years kindof reflects that. Programs are no more than smoke and mirrors to them.
Sounds like AI needs a childhood.
More like AI needs a mentor to provide guidance (metadata) and conditions (restrictions) to shake of bias inferred from the training dataset but still maintain reasonable behavior.
@teetee tee Yes, a childhood.
That’s what they’re doing in the new season of Sword art online lol. They create a simulation and let the AI grow up inside it from a baby all the way to an adult, then they have engrained good morales and common sense, they then extract it and use it for a huge variety of applications (general intelligence ai) pretty crazy premise for an anime show ha
People put it at work before childhood. No wonder it will try to kill us.
So AI is like the ultimate dad joke creator..
Electric nose picker. Dad Christmas present!
Hey I’m sorry I’m not cold I will be home by five but I will see ya tomorrow night love you too love you baby have a wonderful day today and I will talk to him later love him love you and see you soon bye bye man I’ll see him soon love him love you too love you
You want to get rid of rads, so you put a reward for rat tails, and then people make rat farms to harvest the tails to claim the reward.
I’d say working with AI has more in common with working with people than this video might suggest.
Alfonso J. Ramos “The cobra effect” Yeah, when it comes to exploiting loopholes for monetary gain, people will act exactly like AI.
Also known as the “basic problems of a controlled economy”
Alfonso J. Ramos This is true, we do train the AI the same way our brains are trained, so it’s expected.
We already have have emergent superhuman AI in corporations…
I understand and agree with the YouTube recommendations – clicks and views, at the expense of quality content.
They definitely painted the walls of my high school with “suffer”.
XD
This might explain our foreign policy. It must run by AI.
That depends on who’s foreign policy you are referring to.
Sweet Lord in heaven, this thing is evolving with a little help from its designer. !!!!????
None of these talks about AI ever touch on the actual danger in AI development. The corporate seizing of that technology and the development of a permanent super-monopoly that, for all intents and purposes, takes absolute control of the planet. What AI will develop first and foremost is a new weapon in the brand vs brand war. A weapon of mass manipulation so profoundly absolute in it’s execution that the people falling victim will welcome their new chains with open arms, and eager hearts. Humans are extremely easy to manipulate at a 1 on 1 level. Not so much in a crowd. But the true and immediate danger of AI is that it can supercharge targeted delivery and dissemination to a point where billions of people can simultaneously be manipulated at that 1 on 1 level. Granting whoever is delivering that personalized data ABSOLUTE power.
The AI of the very near future will provide an individual with information and advertising individually tagged and developed to suit their digital profile. This destroys the only defense we currently have as a species from absolute tyranny. It breaks down the only barrier that has protected us since the time our ancestors climbed down from the trees. Herd Protection. It will isolate every human being into an echo chamber of branded messages, be they commercial or political. The crowd will no longer exist. And the power that each human being has over their existence will vanish as well.
The true and actual threat AI presents us with is the amplification of psychopathic human en-devour. It will present extremely wealthy psychopathic hoarders with the ability to develop and maintain dynastic control over the breadth of humanity. The Corporate Feudal Age will be the contest that determines who, and how many, sit upon the throne of humanity. And that age is rapidly descending upon us. There is no protection from what is coming. There is no plan. If we continue on our current trajectory, this outcome is inevitable. And it will sweep across the globe far faster than any human response could ever hope to counter.
AI can only become what we make it. Corporations and Psychopaths can, have, and will continue to develop AI that will enslave humanity and cause harm far greater than Mao, or Stalin, or Hitler, could have ever got together in a room and imagined.
Hi, I understand with what you are saying. But since we live in a dualistic universe that instead of aiming for bad- we use that power for to develop good, such as innovation in education,farming production we teach it all types of benefit for mankind and to deliver proper automation with maximum efficiency in resources and time. Then, we as humans, after finally attaining peace and co operation between all peoples and groups who intend to go into AI by following a creed, a Beneficial Peace charter amongst all AI developers that it can only go into good. Just as the father of medicine, Hippocrates, Developed a creed for all doctors in future to abide by, we are in a new developing age and AI takes power over all spheres of human life and thus we need a creed that everyone abides for Ai more than ever.
In order to achieve this level of co operation, we need to have peace and to attain peace groups developing AI need not have greed because what they are co-operating to flourishes everyone. This is what religion wants, fruits of peace. The charter should act like a measurement of how developers deviate against the positive steps and aims the charter recommends to Develop AI. Once it is close to fully being realised and achieved, we can all be nice to eachother and try and conquer space-only if,we are unified through peace and power of Ai.
So AI is like the ultimate writing prompt for “The Monkey’s Paw” stories.
Read the title as “the danger of AL” and was trying to figure out how Al Yankovic had become such a dangerous man
I confess, I may have painted a room in Turdly.
Asks an AI to come up with names without teaching language to the AI.
Axiom of computer programming: Garbage in equals Garbage out.
Sounds like she was talking about neural networks and machine learning.