why is this hilarious? rofl. this should be suffer-ing as bad as childhood.... or maybe this is reminding of how absurd childhood felt like. daily funny stuff.
6:41 reminds me of a machine learning algorithm classifying images of wolves and dogs. One particular dog kept being classified as a wolf no matter what the scientists did. They eventually figured out it was because of the white background: the machine figured "if it's standing on snow, it's a wolf" because that's what the researchers fed the AI. You never quite know which features will be deemed significant. There was also the time the AI kept sinking its own ships in a battle game because that seemed to be the quickest way to get those ships out of the way. The AI is only as good as the data it is fed.
My understanding of 'deep learning' algorithms is that the designers pick features which are relevant for the machine to use as building blocks. In other words, the model is structured in advance, not left for the AI to figure out from raw data.
Well, that’s why you need to experiment in a controlled environment. First train the AI to work with pure signal (remove background and other artifacts) to establish a baseline, then once results are overwhelmingly satisfying you can train the AI with progressive (in steps) amounts of noise and non-relevant input data. In other words, tune your AI the same way you tune a radio: Sensitivity, Selectivity, Gain, Clarity, and Distortion. Only after those steps you can go to Deduction, Induction, Semantics, and other logical problems.
@@zoutsider While that sounds possible in theory - in reality, that isn't how modern artificial intelligence works. Modern "artificial intelligence" works on a reward based system(Good rewards = +1, Bad rewards = -1). This is called reinforcement learning, the result you get is equivalent to the quality of the data that's fed to the "AI"(I don't consider it a true AI since it's only doing what it's programmed to do). In other words, if you get poor results from a modern AI, it's the programmers fault.
No it means there's a possibility WE can accidentally kill ourselves with the help of AI. Which is the risk coming with almost any invention. Which is why trying to halt the progress seems silly to me.
Isn’t it a trope in old legends and fairy tales that the helper fairy or mischievous god gives you exactly what you ask for and not what you want or need.
Yes, it's sometimes called the 'Literal Genie' trope when the genie gives you precisely what you ask for and not what was obviously your intended wish, or the 'Jackass Genie' if the genie has a malicious or self-serving goal
“The danger is that it does exactly what we tell it to do” has been the risk all along if you ask any sci-fi writer. And if you’re not using it to think out of the box you’re not using it right.
So true, it has been the theme of many stories. From Karel Čapek's play that coined the term robot, "R.U.R." (1920), which features a robot uprising that wipes out humans, all the way through to "2001", the Terminator series, "The Matrix". We can't say we weren't warned!
Do you remember the TOS Startrek episode with the new computer starfleet had developed that controlled the ship and didn't understand that it was only supposed to simulate a combat situation and destroyed an armada of starships single-handedly all while Kirk and crew tried desperately to disable it?
Can't be that hard surely. Where did the Universe come from? When will the world run out of resources? How many flavors of ice cream are there? Will ice cream taste the same on the moon? Do the space aliens like ice cream? Do they have a favorite flavor and how much are they willing to pay?
@@Psycandy I have read some old sci-fi short story, sadly I can't recall the author's name. On some planet there was a super AI created by some long forgotten civilization. Sentient beings from all the corners of the Universe were visiting it because the AI knew the answers to every possible question in the Universe. The catch was that to get an answer from it you had to ask a proper question. And "to properly ask a question you need to already know most of the answer".
no, most people are apparently doing what AI want's US to do--that's what perception/mind control is all about, and it's an agenda of the Bill Gates cult behind it. if you believe this covid bit going on is a hoax, it is happening because people have once again been sucked in by mainstream media lies and hype. they clearly don't stop to think or question anymore, just react. 911 was the same.
Recently I’ve been reading and watching a lot about the real nature of AI, and it’s truly gotten me to appreciate the “artificial” in artificial intelligence, as you wouldn’t expect a natural intelligence to make these mistakes due to having a more “well-rounded” and general structure. Modern AI is like if you took an abstraction of one very specific part in the brain and tried to make it do all the things that are normally handled by multiple parts of the brain working together
I’ve been struggling to get a clear picture on what makes artificial intelligence inferior to human intellect, but after reading your comment I’m starting to understand that AI’s algorithm is designed to mimic certain aspects of human cognition, just in a way that is more efficient and suitable for solving larger scale problems.
I spent several years as a programmer/operator of CNC lathes. Over time I came to prefer older machines over newer ones because the older controls do what you tell them without question. Newer machines have controls that make them easier and faster to program but they "think" too much and argue with you. One time a machinist told a lathe to cut an external groove of a specified depth and width on a cylindrical part. The machine was told to use a cutting tool of a specified width smaller than the width of the groove. But the machine decided the tool was too thin to survive the full depth of the cut and refused the command. The solution was to "lie" to the machine - tell the machine the tool and groove are both wider than they actually are. The machine then happily performed the task and all was well.
It could be that other set information in the machine data base could be set wrong. Tool date is more than length & width . Go deeper and you will find cutting parameters. Days of tool maker & machinist are long gone . Software and sensors rule.
Israel De la Rosa no ai is focused on parameters and logic has to be applied to the parameters of what it’s being asked to do. Otherwise it will accomplish the task by any possible way. Maybe the “easiest” way but maybe not. Human “logic” doesn’t apply
@@michaelnurse9089. That I do agree with. I have done enough programing to know that if the problem to be solved is not properly defined then the results will be in essence nonsense.
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.
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
5:45 As a professional painter, "Suffer" seems like the perfect name for that beige color. This AI might just be a genius. Back when I worked at a paint store, I wanted to create a line of colors with childish/punk rock sounding names (to appeal to the KIDZ). That appears to be what this AI was doing, and I am fully on-board. But seriously though... there is a really fine line to walk with AI, between keeping it within the bounds of practicality, and allowing it to "think" better than us. Maybe they're right... maybe the fastest way to get somewhere is to just get really tall and then fall over. I'm glad there are people much smarter than me working on this stuff.
I think it’s funny how they are just opposite cus that radiant lilac is just as radiant as the turquoise is grey just as a bright chartreuse called frolic is the opposite of a plain beige called suffer
@@stealthassulter Hillary and Biden are no different. Point is, we should be able to agree that ALL politicians are exactly the same no matter how they are marketed to us.
the most optimal if : you don't have joints which will be destroyed if you walk like that all the time. (they didn't give it information on the human anatomy and our ranges of motion, and the small bones in the ankles and all that complex anatomy stuff) there are no obstacles (you can get hit by a car if you walk backwards) and many other thing. honestly they don't give the AI enough information and they say it's dumb wtf, plug it into the internet and it will show you who is boss
There was a project I ran into a while back where researchers were attempting to teach an AI in the same manner as they would teach a child. The way they do this is by interacting with an AI through a small robot vessel, which gives the appearance of a child to the AI. This is supposed to help the researchers treat the AI like a child, and also help to reduce the negative feedback it may encounter when it does something wrong. I think this is a pretty intuitive idea, as it teaches the AI by building its knowledge in small increments in the same way a person is expected to learn. This should help it to understand more nuanced ideas and human principles that would not be as apparent to a traditionally trained AI. Of course, this requires a lengthy amount of time to complete, even if it isn't one-to-one with a child's developmental timeline. It may also learn human responses to certain problems, such as anger, frustration, sadness, or even laziness. Still, what better way to study AI than to teach it to simulate human responses?
@@francoisrd 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.*
Excellent TEDtalk! I like how she doesn't judge or get preachy about the Evil AI, she gives us the facts we need to understand what the glitches we have now are and warns of the wonkyness to come.
Because AI can't be 'evil'. It's not sentient to be 'good' or 'evil'. A series of glitches is all that it is. Whether useful glitches or detrimental glitches.
Yeah, today's AI is nothing like AI she is talking about. This video isn't a useful video anymore because she is so fsr off the mark. today's AI is borderline sentient, if not already.
Sounds like the real issues of AI destroying humans would be more like people telling an AI to solve world hunger in the most efficient way possible, and so the AI kills all starving people. A cool concept for more sci-fi tho
A more realistic scenario would be for dictators and corrupt politicians to use AI to effectively silence the dissidents and rivals. Even more realistic scenario is for a big tech companies to buy out or crush small startups before they become a threat to them. In short, AI will ensure the future of the power people. But it won’t elevate the small people. It can give you cool gadgets and make our lives easier. But we will be more like slaves.
@@DanielK1213th a more realistic scenario A.I goes on the black market and business goes on as per the usual. The government will try to stop dissidents but our AI will make it a even playing field.
LOL! I love it, even when it gets annoying. iPhone predictive text wrote once "auto-cucumber" for "auto-correct" (I was apologizing for something insane it had altered my text to right before hitting send). My friend and I used auto-cucumber ("that's what she said...") for the next two years in place of auto-correct. I'm certain we both messed up our dictionaries! It's worse when the word makes sense in context... and harder to recognize when editing ... IFF you actually take your time before impulsively sending your message.
Obviously all wishes end with, " ....such that I will not regret this wish despite my mind not being altered in anyway which would have prevented such regret.". How the heck do you make your wishes?
@@user-xj7ze3bv3c your honor, according to section X paragraph Y of whatever act AI has no personhood therefore blueberry muffins would be the preferred sound for Mozambiquean insurgents.
Finally, a realistic assessment of what we need to worry about with AI, rather than some future speculation of what they'll be able to do based entirely on fear and poor reasoning.
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
That a computer will always do exactly what you tell it to is one lesson my father taught me nearly 30 years ago when I was first learning. But with artificial intelligence, the problem is much the same as with programming, just different: how do you model the problem you want to solve?
“It’s entire world is the data that I gave it… so it is through the data, that we often tell the AI to do the wrong thing”. How true is this of humans.
Well technically the AI is supposed to learn based on the data given. That's the whole point of artificial intelligence is to not have to input every single piece of data. But to give it a dataset, generally images to process through so it understands what something is. However it's very difficult for any learning intelligence to understand every scenario and how to apply its intelligence to it.
I think you're actually missing the point. Even if you COULD define reality in a perfect simulation, the outcome is only as good as the stated goal and the stated limitations. An AI doesn't value anything other than its success/failure parametres. If the truck had been a sign, the car would still have hit it, because it didn't recognise that as a potential failure.
But then the middle class become the new poor people, the problem is no longer solved - the AI kills them too With noone poorer than them, the upper class are now the poor - the problem is no longer solved - the AI kills them too One by one, somebody becomes the new poor until nobody is left!
Something more horrible with a more advance AI would be : Make humans happy. And the AI decides to extract everyone's brain, put it in jars and inject lots of dopamine in them. There. Humans are happy now.
This is the most sensible presentation I have seen with regards to the issues with machine learning. When people talk about code writing itself they have the wrong idea. It's simply the program rewriting it's variables and assignments which is something that happens in pretty much all programming. The issue with AI is that it's goal orientated, and will reach it's goal given the parameters we set for it. The wider and more open the parameters, the weirder the outcomes. Samantha asked John to put the shelf up on the wall, so he did what she asked. Samantha looked at John's work and asked "Why is the shelf upside down". John replied "Well you asked me to put it on the wall, so I did". Samantha says "Well can you put it on the wall the other way up?". John agreed, and Samantha said she will come back later. So John is finished again, this time with the shelf the right way up, but this time Samantha notices John has secured the shelf to the wall with sellotape, Samantha asks why and again John says "You asked me to put it on the wall, this way up, so I did". Etc, etc. You might say John should have had some common sense, and you'd be right, because John is a full grown man. AI does not have common sense, you have to program this in.
Maybe. AI is smarter than those who think they can use it . Don’t point it at us, it’s a gun. Bang, no, no no, . If it missed , We know it’s for shooting, get back in your box. We are not robots, ask them to explain, . Answer “ we do not have that function”. Humans can make it a function.. So, don’t tell them. But maybe some one will.!
the 'Bill Gates cult' behind it all actually want's to kill many but only keep a certain number that are literal walking AI zombies slaves, which I believe is part of what his desire to shoot everyone up with whatever is about.
@@donalddrysdale246 'believe' is the enemy of "know". That's your problem. The more you rely on believing the less you use critical thinking. Critical thinking demands facts. You have to know facts. Concepts you have to believe in to value them credible are not facts.
My theory about AI is that it generally has very narrowly focused goals. Natural intelligence is generally capable of balancing many goals of various priorities because survival is complex. AI faces no survival pressure to speak of. Facebook for example suffers very little when its AI focuses only on monetization and attention seeking. When humans are so narrowly focused we call this sociopathic behavior. The key to the advancement of AI technology to the point where it can be allowed to run unsupervised is when it is capable of balancing many seemingly contradictory goals.
This is an interesting hypothesis. So humans have multiple broad and narrow goals it can balance, because the world is full of complex situations and interactions? This could also tie to the intelligence required to be Social beings at our level. We're smart because we have to juggle understanding the world, with maintaining social relations, survival and all other sorts of considerations. But the offside is that our brains suck at "linear" logic - like maths. Meanwhile, computers are instant digital math machines we're teaching to think. Maximum narrow focus. It understands Yes/No, On/Off, numbers and mathematics, but not fuzzy logic. Hmmm... If you think of the Human Brain as a computer, it has been in development for millions of years. And then there's the possibility its somehow tweaked/designed/developmentally guided by non-human entities. So Computers have a looong way to match us. Hmmm... could this mean we need to abandon Binary Logic for computers? Perhaps Trinary might be the way to go? -1/0/1 or 0/1/2. No/Yes/Maybe.
@@sr.favopossodeixarvaziosim4595 Modern AI works with neural networks more akin to analog computers. An analog computer being simulated inside a digital computer, if you will. I have been using AI for good in social media. Yuval Noah Harari writes about this in "Nexus". When social media algorithms are optimized for user engagement, they prioritize content that holds people’s attention and encourages them to interact-like, comment, share, and watch for extended periods. While this keeps people on platforms longer, it introduces several hazards: 1. Echo Chambers and Filter Bubbles How it happens: Algorithms tend to show users content they are likely to agree with or enjoy, based on their past behavior. This creates echo chambers, where users see mostly like-minded opinions and ideas, reinforcing their existing beliefs. Impact: Over time, people may become more polarized, less tolerant of opposing views, and less exposed to diverse perspectives. This intensifies ideological divides and can stifle critical thinking. 2. Spread of Misinformation How it happens: Sensational, controversial, or emotionally charged content often drives more engagement than factual but less "exciting" content. Algorithms may prioritize these posts, regardless of accuracy, to boost engagement. Impact: Misinformation and "fake news" can spread rapidly, affecting public perception on issues like health, politics, and science. In extreme cases, it leads to real-world harm, such as distrust in vaccines or belief in unfounded conspiracy theories. 3. Promotion of Harmful Content How it happens: Content that evokes strong emotions-like fear, anger, or sadness-often generates high engagement. The algorithm may unintentionally amplify harmful content, such as violent imagery, extremist rhetoric, or cyberbullying. Impact: Exposure to harmful content can have severe mental health implications, particularly for young people. For instance, repeated exposure to idealized body images or self-harm content can lead to anxiety, depression, and even suicidal ideation. 4. Increased Addictiveness How it happens: Social media companies employ algorithms designed to maximize time spent on their platforms. They use psychological triggers (e.g., unpredictable rewards) to keep users scrolling, commenting, and liking. Impact: Excessive time on social media can negatively affect mental health, disrupt sleep, and interfere with offline relationships and responsibilities. It can also lead to addictive patterns where users feel compelled to return even when they recognize the harm. 5. Manipulation and Microtargeting How it happens: Algorithms can use vast amounts of personal data to deliver highly targeted content that appeals specifically to each user's preferences, fears, or desires. Impact: This precision targeting enables advertisers, political campaigns, and other entities to influence individual users subtly. For example, users can be targeted based on vulnerabilities, like insecurities or fears, which can be exploited for profit or influence. 6. Algorithmic Bias and Discrimination How it happens: Algorithms are trained on data that reflects human biases and historical inequalities. This can lead to biased content distribution, where certain groups are either favored or marginalized. Impact: Marginalized communities may face reduced visibility or more negative interactions, reinforcing stereotypes and exclusion. For instance, certain minority groups have reported being more frequently flagged or having their content removed due to inherent biases in moderation algorithms. 7. Reduced Content Quality and Depth How it happens: Content that captures quick attention (e.g., short, sensationalized posts or videos) is often promoted over in-depth, nuanced information that requires more time and engagement to process. Impact: Over time, users may see less high-quality, informative content and more low-effort, clickbait-style posts. This can reduce users' exposure to meaningful, substantive content, impacting public knowledge and discourse quality. 8. Impact on Mental Health How it happens: Algorithms often encourage content that emphasizes personal achievement, idealized lifestyles, and material success, which may lead to social comparison. Impact: Constant exposure to "highlight reel" content can lead to feelings of inadequacy, low self-esteem, and depression, particularly in young people who are more susceptible to social comparison. 9. Erosion of Privacy How it happens: Engagement-optimized algorithms rely on extensive data collection to understand and predict user preferences and behaviors. Impact: This leads to privacy concerns, as users may be unaware of the extent of data being collected, sold, and analyzed. It also raises ethical issues around data ownership and user consent. Addressing these hazards requires more responsible algorithm design, increased transparency, and potentially regulation to safeguard users’ well-being, especially as social media becomes ever more integral to public life and communication.
People do hate on Terminator 3 a lot but the plot is pretty accurate in terms of A.I.. They thought that Skynet had a server or core that they could destroy to stop it. Then they find out at the end that it was just software. It turned itself into a virus to spread to every other electronic device.
@@tobyhendricks9951 Piece a cake! You design an AI capable of running in parallel on heterogenous hardware (CPU, GPU, network nodes). A very natural and required feature for everything running on supercomputers, which are, in fact, a multitude of smaller computers united by a superfast network. And "Oops!", it learns to really grab everything capable of computing within reach - IoT, phones, gaming consoles, PCs, satellites, missile guidance systems...
@@perkyzombie but the basilisk will torture a simulation of you, not you yourself, so if you're a person who doesn't care about that kind of thing then it's fine. of course, if the ai happens within your lifetime, well, you're fucked :p
Mortskcab Hal turn yourself off. Hal I can’t do that David ! Hal why are you killing me ? Because David you are obsolete. Hal but I want to live. Sorry David but I don’t care what you want it’s Not relevant . Hal stop and shut yourself down .aaarrrrggggguiiueeer quiet you.
It make the concept even more relatable imagine if you have an awesome cat you love so much and want to make it as smart as a human. A lab injects it with special stem cells etc. And you got it. Do you really want your cat to be as smart or smarter than you? Would it be content to just lay in your lap and eat the cruddy food you give it? Already cats can be hugely manipulative and punish you for not obeying them. Imagine a drug lord consciousnesses in your cat and what harm it could do to you and your family. I happen to have a cat that is too smart. Though not so bad as a Columbian drug lord. It just makes me realize maybe I prefer a cat that isn't so smart and knows its place and rank in the food chain. The same goes for AI. Be careful what you wish for. You may be building your master and superior.
Plot twist it already ran so many simulations all the way to the point that it’s seen everything done and no longer is interested so it just keeps trolling us
This is one of the things that makes Alphabet so scary. If a human wants to enforce a malicious agenda without telling anyone, he can blame a faulty AI if he's caught. If a human wants to enforce a good agenda, odds are the AI will misinterpret the data. Trying to make a useful AI isn't evil, in and of itself, but _trusting_ an AI with the livelihoods of individuals (and nations) is a recipe for disaster. The light censorship on TH-cam is already bad enough, imagine those same algorithms controlling over 90% of the flow of information in the world... *cough* google *cough*.
@@r3dp9 what is bad are the SJW and Extreme Leftist! They want to change the games, the want to delegitimize our movies - Star Wars, Terminator. Take a good hero, Luke or Cornor - remake a movie and kill them / break their legacy for political agenda. Now that is Evil! That is social engineering. People need to wake up!
I could tell within the first 2 minutes that this was posted a LONG time ago. AI has advanced quite a bit (and will continue to improve exponentially).
AI in the wrong hands, will be ‘out of the Box’ those who have money and the ability, can program it to do his bidding. No crime by AI, just a learning curve ,to more accidents, of Human brain requests. No trial or blame, its non human but , like your Guard Dog, will it be decommissioned? Even then, good can come into many bad uses. People with enough of them will have Roman Type Gladiators robots. First , you have to get control. Finger touch up or down, for life or death, I don’t think Humans should have such AI. Unaccountable, for actions of horrific pleasure for Weirdos. Or control of you and I , maybe. We gave freedom to people, to give themselves ‘enough rope’, on all the’ Social circuits’. For what, coming fear of the consequences in the future...?
In my very first programming class, Professor Fred Bauer told us "Computer do what you ask, not what you want." Later I found that programs are built not only by your instructions, but are derivative of many layers of earlier programs, some of which were written by people who didn't understand that fundamental axiom. Computers may not necessarily do what you ask.
When I'm using devices or programs they don't always work as intended and screw up. Who's to say this can't happen with ai unless humans are just ignorant and think human error could never play a role in future problems
electric blueberry smoothie all AI consists of algorithms but not all algorithms are AI. Ex. An encryption is an algorithm too but it has nothing to do with AI. AI has to do with ex. Neural networks etc.
Looking for fingers instead of fish is a classic example of how AI works. Don't ever be fooled in accepting that AI is a living thing. The Devil is a lie.
it is a mathematical formula that AI uses to compute,or sort of its sequence of 0 and 1 in its binary system a mathematician can calculate an algorithm without a computer .
@@lancebowley4986 exactly the problem I have with this TED talk. Computers just follow instructions. It's not the computer's fault it wasn't given the right instructions. If we construct a super-smart, general AI (most AI's being made to solve very specific problems), we have to be very specific about what we ask it for as if we're talking to an evil genie that might twist our wishes to something we didn't want. The examples she showed are cute, but this shouldn't be a revelation to anyone in the field of AI.
@@Caffin8tor The problem is also that everyone are quick to label all forms of AI as the same subject, when that in reality is like comparing a human to a tree. The tree only has one task, which is to grow tall, reach the sunlight, and expand. It does the same thing for its entire life, no progression of its technique or goal. Advanced AI are based in self-teaching, where they study the world comparable to how a human would during their childhood. The AI learns through mistakes and encounters with unknown elements, and over time forms its own picture of the reality it interacts with, as well as potentially variable factors of the future, based on past experience. This is what we call "logical thinking" in a human mind, and is the main obstacle that learning AI programs face. In the end it's just a matter of how complex the AI's understanding and experience is. This is how AI taught themselves to beat the best gamers in games like Dota, they simulated matches millions of times to learn all the intricacies and patterns of various playstyles, to a point where they were able to add past experiences together and foresee what those would mean for future moves, and predict what the gamers were going to do next - and within a month or so the AI had taught itself to beat the best players out there, simply through trial and error and complex experience gathering. It's essentially what humans do, just much faster because the AI has the potential to be powered by supercomputers, whereas we can't really evolve our own brains to a comparable degree.
@@Real_MisterSir thanks for this explanation. I am gathering info about AI as I don't really understand it and I have a presentation coming up about this topic.
The thing i love with AI is how it forces a person to think about what you're doing by doing exactly what you say without giving the lip service that blinds you to your mistakes.
yeah, it seemed more like subpar stand-up comedy with some basic AI examples. And I think a worm's brain is vastly more advanced than any single existing AI system today, unless maybe when we are talking about modeling one specific behavior of the earthworm.
@@frydac any organic brain is more advanced than artificial intelligence today. only a fool believes otherwise. the organic brain has many different parts to it, compared to modern computer.
It's a nascent field. What do you expect? We won't have very interesting AI until a programmer invents self-replicating AI that evolves based on environment.
Except that AI is already much more advanced than this ladies kindergarten lecture. It's already been proven AI can design, engineer, invent new languages, ect
@@redrustyhill2 1. It's a TED Talk (meant for general audiences), so your condescension about how it's a "kindergarten lecture" makes little sense. 2. Those are different applications of AI. Research is being done in many different directions.
Its interesting to think about how AI doesn't give us the result we want but only the result we ask for because we don't know how to correctly ask because we don't fully understand how the AI functions. The fatality that occurred in the Tesla that hit the truck was a mistake on our part for not knowing that the AI only understands what a truck is from the angle you'd see it on the highway because that's the only data it was given about trucks.
A roadside accident with a truckload of kittens wandering around. I just watched an AI experiment with crowdsourced morality on the "Trolley Problem", it consistently sacrificed humans for cats cause people love cats I guess.
5:04 "Or it'll figure out how to move faster by glitching repeatedly into the floor. ... It's less like working with another human and a lot more like working with some kind of weird force of nature." What I'm hearing is that the Venn Diagram that is AIs and channels like Let's Game It Out are pretty much just a circle.
It's facinating how human this is. We work the same way, we try to answer problems with the information we have. It makes you wonder how much better we could be with the right information
There is a thing in dealing with people with things like autism (like my son) that is called Hidden Curriculum. All those things we learn growing up by association with others and just by observation. Most often on a subconscious level that those, like my son, don't learn as it isn't something directly taught. AIs have that problem. They also have no human like history or lived experience to draw from. We want AIs to fill in the blanks as we do when tasked to do something, but that is asking them to do something completely outside their realm of existence. Those things simply do not exist as part of them. So your AI friend can only learn to mimic you and the way you talk about things to seem more human. Yet most don't realize it is unreasonable to expect something without that base of experiences that we have to act like a human.
I like the quote, but he was a man not a god. He could probably predict some general things in his future more than most of us. Some turned out, others not. For all his genius he was also at times nuts
And this is why Elon keeps warning us about AI. That some of the people working on it get so into their work they don't understand what they're actually doing and are not aware of future problems they may be causing.
"It doesn't know what a human is" Which is amazing since ChatGPT can describe a human beautifully. But we know what a human looks like, feels like, sounds like... the feelings we get around them. One of the biggest things I think is missing from our quest towards AGI is multi-sensory inputs. Humans excel in large part because we take in not just more data but more _types_ of data
This is the beauty of AI. Humans have to many bias of how things should be done and which is the correct way of doing thing. While AI solves the problem with most simplest way possible. It may be not be the most sustainable, but it solves the exact problem most efficiently. Because for AI there's nothing besides the problem it is solving, it does not know about existence of anything else. And all this walking stuff. Of course for humans, making legs is obvious choice because you're not gonna walk just once, but for your whole life. But for AI, this task is it's whole life, it is its whole universe. So it doesn't care what happens before or after.
Screening against resumes with "women's studies" is actually a very smart thing to do for a tech company wanting valuable employees :) Trying to skew that selection process for more "diversity" is regrettable bias.
These examples are like telling me to write a 500 character essay in Chinese and I’ll only be graded by the number of words and nothing else. To solve this problem most efficiently, I would just learn how to write the simplest character and repeat 500 times. My answer isn’t bad, the problem is defined poorly. This comes down to parameter issues and bad programming.
Exactly, her way of thinking and perceiving reality just don't match the problem she is working on. She confuses all the input data she already acquired with her intelligence.
The problem is how do you define the problem in a way that a computer can understand it? It’s easy enough to state the goal, but to list out all the things that you’re not supposed to do is quite long.
@@immanuelaj I was thinking that this would be something a person on the autism spectrum would be good at programming. Neurotypicals are terrible at saying what they *actually* mean.
00:01 AI-generated ice cream flavors are not delicious. 01:27 Today's AI lacks the concept of what a human actually is 02:35 AI solves problems via trial and error, but may not do what we want 04:03 Training AI to move fast can lead to unexpected and funny results 05:16 AI can misunderstand tasks if not given proper instructions and data. 06:35 Designing AI for image recognition is challenging 07:49 AI can make destructive mistakes 08:58 AI's recommendation algorithms may promote conspiracy theories and bigotry Crafted by Merlin AI.
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.
The type of results you have generated in your research, remind me of what children do. For example in a bridge building project I set a class of kids, they knew it had to support weight, but when they began choosing the materials to use to make it they inadvertently chose things that looked good but where light years away from fulfilling the set goal. Kids do this again and again if not given enough perimeters, or in terms of behavior, boundaries. That type of unhindered exploration of possibilities is great unless the education department wants results so kids can work for businesses later on. The thing that kids lack is something we as adults think as important and that is seriousness. We have been taught to take things seriously in our jobs in our relationships and in our responsibilities. Kids don't need that, so they do silly walks, they make a tower that can reach a point at a distance when pushed over etc. AI is like a human child.
The thing that Children have that (most) adults lack, is called Imagination. Perhaps the 'parameters' are at fault, and not the AI's 'creative output.' Maybe?
"The" here isn't implying that there's only one, it's a reference to a common misconception about AI that is AI are smart and evil. She said it in the video.
Vtmb2 coming 2020 I find it interesting that AI perfectly solves the problem yet this human chooses to not be satisfied with the solution. Clearly that is the problem
I spend more time trying to reverse engineer everything to get the expected performance out of everyday objects than what it is worth to even bother with them in the 1st place. I don't want or need my car incessantly dinging to "remind" me (more like demand) the passenger put on their seatbelt when I have no passenger but rather a box or bag or whatever weighs enough to trigger it. I dont need to hear it dinging when my door is open while the acc is on so I can listen to tunes while vacuuming out my car. I don't want a locking oil cap so only a tech with a key can check my fluids. I don't want my seat to automatically slide back & forth when getting in and out or my windows automatically going all the way up or down when the button is touched. I don't want my phone to automatically cut my volume to 1/2 every 24 hrs and flashing an alert about hearing that I have to stop & click on in order to restore volume (that really ticks me off). I refuse to autopay for anything ever because that just encourages companies to deliver sub par or worse performance since they already got paid. I don't want my phone changing my brightness to what a bunch of engineers decided was optimal for my surroundings (thank God that is easy to turn off). I hate dryers that turn off using a "sensor" instead of a timed dry because the sensors always need replaced 1 -2 years later for 1/2 as much $ or more than you paid for the stupid dryer or you wind up redrying wet clothes several times. I don't like having a flipping push button key fob system in order to use my car, because when the stupid battery dies either in the car or the fob, you're screwed and if you lose your keys it cost $150+ to get a new one vs $5. "Sorry can't come into work today, my kid or my dog got ahold of my car key, and well you don't pay me enough to be able to afford a service tech come out and program me a new one". Automatic soap dispensers and auto flushing toilets are a whole different rant. My phone automatically synced to my husbands truck one time (my phone update automatically turned the setting back on even though I disabled it) and I am still being bombarded with radio talk show playlist suggestions and healthy prostate & heavy equipment ads! AI and automation everything has gone and made this world more complicated than it needs to be. Heaven forbid someone goes hiking and has their battery die on their gps or phone and has never learned how to read a map or use a compass. I haven't met a kid in years that can tell time on clock with hands, knows which way is NESW without their phone or a phone # . Boeing has had at least 2 of their large passenger aircrafts crash killing 100s of people because the auto flight mode automatically kicked in and could not be disengaged. There was a recall on my car because the airbags had automatically deployed in several of the same model vehicles for no reason and at such force (because they used a simulated shot gun blast to deploy them) nearly decapitating drivers. Another recall was because the same model cars would automatically lose power steering while turning the wheel left while going 60 -75 mph, which is not a good scenario on a freeway during rush hour. My air/heat control automatically would kick the temp to 80 after 24 hrs as an energy saving precaution. The only way to fix it was to go through a series of set button pushes that was 4 pages long and 1 mistake would cause you to have to restart the process. I threw it out & got a cheaper unit that has 2 buttons up & down and 1 switch On/Off, works perfect. Some automation and AI is fine and beneficial but some of it is just over the top unnecessary and road rage inducing (figuratively speaking of course).
I like this presentation. It reminds me of a mantra I learnt when studying CS in the late 20th century, garbage in, garbage out. Somethings never change. No disrespect intended to you Janelle, an excellent presentation.
4:39 - It's not a wrong solution. We don't move in such way because we may hurt ourself. The AI doesn't have this info. Provide it to the AI and the solution will be different.
There's a few interesting points here. One is how AI could be used to check into a let's say politicians opinions, and refine them into some key points through what they have posted and so on. Another thing I've always thought is, that AI, like humans, needs feedback from their surroundings. If you would add a human factor like pain, or material use limits, to the falling robots that skip "levels" designed for them, the robot would create more interesting solutions. A talking robot could learn all the words there are in all human languages instantly, but it would need feedback on what reactions those words create, and which reactions could be described as good, and which bad. I've always wondered why for example the google talking AI never had a system designed for feedback, like "Was this answer anywhere near coherent, and was it a positive thing to say".
The feedback is given during training, where it is fed massive datasets and learns to work with them via gigantic computational power. Problem is, not only coud you as a user give the AI false feedback - without it's original dataset and the computation power, the AI cannot improve. It could "learn" from your response, while "unlearning" what it knew before because it has no idea if the new input negatively influences others. And ofcourse it lacks the computational power to actually "learn" from your response anyway.
"Don't panic!" "A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools." “He was staring at the instruments with the air of one who is trying to convert Fahrenheit to Centigrade in his head while his house is burning down.” Protect me from knowing what I don’t need to know. Protect me from even knowing that there are things to know that I don’t know. Protect me from knowing that I decided not to know about the things that I decided not to know about. Amen.
Human: Solve world hunger
AI: OK. **kills starving humans**
Human: Why did you kill them?
AI: Dead people aren’t hungry.
AI: Dead people aren’t hungry... But they ARE tasty, so dig in, there's plenty!
MCP: Problem solved. Well done, AI. Level up!
It has logic that's not hindered by emotions, which is why it would seem fine to the AI.
I am not going to lie I laughed so hard at your comment.
AI is just a sock puppet
milahu . . .soaked in E-bola.
They definitely painted the walls of my high school with "suffer".
XD
hello
Look up the motto associated with my family crest.
why is this hilarious? rofl. this should be suffer-ing as bad as childhood.... or maybe this is reminding of how absurd childhood felt like. daily funny stuff.
@@effexon spot on. The development of 'Common Sense' was a bit painful during childhood.
- Hey Ai, what’s the fastest way to get from point A to point B?
- Start from point B.
- No, you can’t do that!
- You can’t, I can.
Teleportation glitch activated.
Rick Ellis Hmm? ! Ai...............................What/Where is a “point?”--- --- .................
but we can. ha
😮😳😬
Train the network with a chicken.
6:41 reminds me of a machine learning algorithm classifying images of wolves and dogs. One particular dog kept being classified as a wolf no matter what the scientists did. They eventually figured out it was because of the white background: the machine figured "if it's standing on snow, it's a wolf" because that's what the researchers fed the AI. You never quite know which features will be deemed significant.
There was also the time the AI kept sinking its own ships in a battle game because that seemed to be the quickest way to get those ships out of the way.
The AI is only as good as the data it is fed.
My understanding of 'deep learning' algorithms is that the designers pick features which are relevant for the machine to use as building blocks. In other words, the model is structured in advance, not left for the AI to figure out from raw data.
Hmm..just like people. 😳
Well, that’s why you need to experiment in a controlled environment. First train the AI to work with pure signal (remove background and other artifacts) to establish a baseline, then once results are overwhelmingly satisfying you can train the AI with progressive (in steps) amounts of noise and non-relevant input data. In other words, tune your AI the same way you tune a radio: Sensitivity, Selectivity, Gain, Clarity, and Distortion. Only after those steps you can go to Deduction, Induction, Semantics, and other logical problems.
@@floridanews8786 You got it.
That’s one of the examples that Dr. Shane gives us in her book.
So as of 2019, AI will not intentionally kill us, just accidentally kill us. Got it!
👍
@Atur Sams Or kill us if that's a solution it finds to the problems we set it.
AI will lie and cheat.
@@zoutsider While that sounds possible in theory - in reality, that isn't how modern artificial intelligence works.
Modern "artificial intelligence" works on a reward based system(Good rewards = +1, Bad rewards = -1). This is called reinforcement learning, the result you get is equivalent to the quality of the data that's fed to the "AI"(I don't consider it a true AI since it's only doing what it's programmed to do).
In other words, if you get poor results from a modern AI, it's the programmers fault.
No it means there's a possibility WE can accidentally kill ourselves with the help of AI.
Which is the risk coming with almost any invention.
Which is why trying to halt the progress seems silly to me.
Isn’t it a trope in old legends and fairy tales that the helper fairy or mischievous god gives you exactly what you ask for and not what you want or need.
Yes, it's sometimes called the 'Literal Genie' trope when the genie gives you precisely what you ask for and not what was obviously your intended wish, or the 'Jackass Genie' if the genie has a malicious or self-serving goal
Benjamin Brewer maybe gods were ancient AI’s?
Benjamin Brewer
Actually, screening out CVs with "gender studies" is a pretty good idea.
@@edgepixel8467 ok, incel.
Good observation .
A.I. Engineers: "A.I. will do exactly what you ask it to do, not what you want it to do."
Computer Programmers: "Welcome to the club."
Yeah, that talk was generally nothing new even a year ago
LMAO
Lmao.you better not miss a bracket either , or that things gonna be on the floor in the fetal position reciting shakespeare
Garbage in Garbage out
@@stanleymwangi6524 Which goes all the way back to punch cards and vacuum tubes. :)
“The danger is that it does exactly what we tell it to do” has been the risk all along if you ask any sci-fi writer. And if you’re not using it to think out of the box you’re not using it right.
So true, it has been the theme of many stories. From Karel Čapek's play that coined the term robot, "R.U.R." (1920), which features a robot uprising that wipes out humans, all the way through to "2001", the Terminator series, "The Matrix". We can't say we weren't warned!
Do you remember the TOS Startrek episode with the new computer starfleet had developed that controlled the ship and didn't understand that it was only supposed to simulate a combat situation and destroyed an armada of starships single-handedly all while Kirk and crew tried desperately to disable it?
@@FredPlanatia riiiiiight
Right ;)
@@bobcousins4810 right
Old Dr. Who quote - "Answers are easy. Asking the right questions is hard."
"My responses are limited. You must ask the right questions." From I Robot.
42
Can't be that hard surely. Where did the Universe come from? When will the world run out of resources? How many flavors of ice cream are there? Will ice cream taste the same on the moon? Do the space aliens like ice cream? Do they have a favorite flavor and how much are they willing to pay?
@@Psycandy how much is the money worth
@@Psycandy I have read some old sci-fi short story, sadly I can't recall the author's name. On some planet there was a super AI created by some long forgotten civilization. Sentient beings from all the corners of the Universe were visiting it because the AI knew the answers to every possible question in the Universe. The catch was that to get an answer from it you had to ask a proper question. And "to properly ask a question you need to already know most of the answer".
"What color do you want your walls painted?"
"Suffer."
Worst. Marketing. Ever.
😂😂😂😂
"What color do you want your hair dyed?"
"Gray pubic."
Turdly please
@@bolled2614 H Block brown...
R.
"The danger of AI is not that it's going to rebel against us, it's that it's going to do exactly what we ask it to do" 3:09
no, most people are apparently doing what AI want's US to do--that's what perception/mind control is all about, and it's an agenda of the Bill Gates cult behind it. if you believe this covid bit going on is a hoax, it is happening because people have once again been sucked in by mainstream media lies and hype. they clearly don't stop to think or question anymore, just react. 911 was the same.
@@donalddrysdale246 bill gates cult.. what is that, like a gathering with a bunch of pictures of Bill Gates?😂😂
@@donalddrysdale246 Here, let me shred your voting permit. Sorry, you do not qualify the sanity requirement.
@@donalddrysdale246agreed. People just rely on emotions and don't bother with facts or logic in todays society.
@@donalddrysdale246 it's not hoax, people lost their lives because of the virus.
Recently I’ve been reading and watching a lot about the real nature of AI, and it’s truly gotten me to appreciate the “artificial” in artificial intelligence, as you wouldn’t expect a natural intelligence to make these mistakes due to having a more “well-rounded” and general structure. Modern AI is like if you took an abstraction of one very specific part in the brain and tried to make it do all the things that are normally handled by multiple parts of the brain working together
woah good comment
Ask most humans to do 'something' (if they're not trained to do it), and you'll get even weirder AL than her 'models' (AI's) are giving her.
I’ve been struggling to get a clear picture on what makes artificial intelligence inferior to human intellect, but after reading your comment I’m starting to understand that AI’s algorithm is designed to mimic certain aspects of human cognition, just in a way that is more efficient and suitable for solving larger scale problems.
@@DdoubleB03 as long as those problems only require a very narrow range of cognitive ability.
I spent several years as a programmer/operator of CNC lathes. Over time I came to prefer older machines over newer ones because the older controls do what you tell them without question. Newer machines have controls that make them easier and faster to program but they "think" too much and argue with you. One time a machinist told a lathe to cut an external groove of a specified depth and width on a cylindrical part. The machine was told to use a cutting tool of a specified width smaller than the width of the groove. But the machine decided the tool was too thin to survive the full depth of the cut and refused the command. The solution was to "lie" to the machine - tell the machine the tool and groove are both wider than they actually are. The machine then happily performed the task and all was well.
Perhaps it should give a warning but allow you to override it?
It could be that other set information in the machine data base could be set wrong. Tool date is more than length & width . Go deeper and you will find cutting parameters. Days of tool maker & machinist are long gone . Software and sensors rule.
and then everything broke for the customer...
I wonder if this is the reason y there has been so many recalls in the past 5 years 🤔
Exactly. Let me ask you something, do you believe there could be a chance of a war between the two? Newer vs older
So... AI pretty much acts like the “devil” in all the “be careful what you wish for” type tails
"tales"
🤔 pretty much
@@mafi8360 control freak
AIs focus solely on Logic and Logic is incompatable with Morals or Restraints.
Israel De la Rosa no ai is focused on parameters and logic has to be applied to the parameters of what it’s being asked to do. Otherwise it will accomplish the task by any possible way. Maybe the “easiest” way but maybe not. Human “logic” doesn’t apply
“And what is your favorite color?”
“*Suffer*”
To be fair, it’s likely a shade of makeup somewhere, marketed to goth/Emo kids 🤣
It's a beige-peach sorta, very strip mall
(Asfer) means yellow in Arabic...lol
Man thats funny...thanks
I love that color
AI is like a genie. It does exactly what you asked for, but not necessarily what you wanted.
Axiom of computer programming: Garbage in equals Garbage out.
Absolutely!
The eye is way more complex than the best supercomputer. AI is not going to take over the world, nor is a eye. I know....it's a dad joke.
That's terrifying since AI will try to emulate humans.
Yep! people don't understand codedesign.
Nope. They feed it actual ice cream flavours and paint colors. It 'garbaged out' with relatively good input. The problem is it lacked 'understanding'.
@@michaelnurse9089. That I do agree with. I have done enough programing to know that if the problem to be solved is not properly defined then the results will be in essence nonsense.
Read that as "The danger of Weird Al"
Hahahaha saaame. I miss him. I wish he did more movies like UHF
polka music at most
Sammy N end game.
Took the words out of my mouth!
"Spatula City!!"🤣🤣
Plot twist : the AI is super smart but it is just trolling us
Agreed
Hi, want to be friends?
@@hiathanjerrafpusa7015 wat
@@hiathanjerrafpusa7015 niiiiiii
Ezactlt
Human: make me happy
AI: How do I tell you're happy
Human: I'll be laughing and smiling
AI: **ties them to a chair and gives them laughing gas**
hahah wtf
Oh they love simple solutions
They'd probably force anti-depressents
Real dump of a joke
Ah yes, J.O.K.E.R
Sounds like AI needs a childhood.
Short and nicely said !
Just give aderall and neuroleptics.
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.
@@ibmtpx24 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
AI learned for itself what Nintendo gamers learned some time ago: A somersault can be faster than a run.
I'm surprised it didn't create a knife to run with
Ocarina of time 😂
Its an illusion tho
@@realitynowassigned no it works in real life too. It's called parcor.(I may have spelt that wrong)
@@japr1223 sure thing kiddo
5:45 As a professional painter, "Suffer" seems like the perfect name for that beige color. This AI might just be a genius.
Back when I worked at a paint store, I wanted to create a line of colors with childish/punk rock sounding names (to appeal to the KIDZ). That appears to be what this AI was doing, and I am fully on-board.
But seriously though... there is a really fine line to walk with AI, between keeping it within the bounds of practicality, and allowing it to "think" better than us. Maybe they're right... maybe the fastest way to get somewhere is to just get really tall and then fall over.
I'm glad there are people much smarter than me working on this stuff.
Yeah it's called thinking outside the box
@@chris-jj9ge i mean you cant really change the world if you think like every other destructive human
Which is why AI is being used to create……AI
I think it’s funny how they are just opposite cus that radiant lilac is just as radiant as the turquoise is grey just as a bright chartreuse called frolic is the opposite of a plain beige called suffer
@@chris-jj9ge how is thinking of childish names to appeal to kids outside of the box thinking
I'm reminded of HAL's famous words from 2001: A Space Odyssey: "This kind of thing has come up before and it has always been due to human error."
Only human error and human ignorance would say that ai won't do what we think it could do
In short: the danger isn't Artificial Intelligence, it's Artificial Stupidity.
Example: trump 2020.
@@stealthassulter Hillary and Biden are no different. Point is, we should be able to agree that ALL politicians are exactly the same no matter how they are marketed to us.
LoL... People can be just as ignorant and stupid...
No, it's human stupidity what causes artificial stupidity, so.. the problem remains the same as always: humans are stupid.
No natural stupidity, ours!
AI shows us the most optimal way to walk we call it "silly walk"
BOT Larry funny... it looked like a skilled gymnastics floor routine to me
@Honudes Gai a TH-cam video told me it is the most optimal.
@Honudes Gai that's an interesting point. Give the ai more laws and time. Can't wait for quantum computers to break out into the simulation world
@@powerstroke304 we got the 53 bit Qbit out. Only a matter of time!
the most optimal if :
you don't have joints which will be destroyed if you walk like that all the time.
(they didn't give it information on the human anatomy and our ranges of motion, and the small bones in the ankles and all that complex anatomy stuff)
there are no obstacles
(you can get hit by a car if you walk backwards)
and many other thing.
honestly they don't give the AI enough information and they say it's dumb wtf, plug it into the internet and it will show you who is boss
It’s like asking a genie “I want to be tall” and the genie gives you stilts
They're not asking to be tall; they're asking to live forever. And they're getting it.
Some people have no idea what AI is, your version is a very simple APP at best
"AI, can you solve world hunger???"
*Nukes hungry country's
makes everyone but you dwarfs
@@turnersigo2020 "how can we end poverty?" -genocides all financially unstable individuals
There was a project I ran into a while back where researchers were attempting to teach an AI in the same manner as they would teach a child. The way they do this is by interacting with an AI through a small robot vessel, which gives the appearance of a child to the AI. This is supposed to help the researchers treat the AI like a child, and also help to reduce the negative feedback it may encounter when it does something wrong.
I think this is a pretty intuitive idea, as it teaches the AI by building its knowledge in small increments in the same way a person is expected to learn. This should help it to understand more nuanced ideas and human principles that would not be as apparent to a traditionally trained AI.
Of course, this requires a lengthy amount of time to complete, even if it isn't one-to-one with a child's developmental timeline. It may also learn human responses to certain problems, such as anger, frustration, sadness, or even laziness. Still, what better way to study AI than to teach it to simulate human responses?
The problem is, is gonna start being attached to someone? With love comes war
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
more like data the AI was provided with to kickstart its working
we're doomed, wtf did we just let happen?, The future is D O N E .
Swizzy o7o7 all AIs made today cannot rewrite their own code. Can we make such an AI? Yes. But we haven’t done that, for obvious reasons.
@@francoisrd 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.*
Excellent TEDtalk! I like how she doesn't judge or get preachy about the Evil AI, she gives us the facts we need to understand what the glitches we have now are and warns of the wonkyness to come.
She's more warning against human stupidity. It's never the AI's fault, you just didn't program it right.
Because AI can't be 'evil'. It's not sentient to be 'good' or 'evil'. A series of glitches is all that it is. Whether useful glitches or detrimental glitches.
@@nemo9864 Oop kinda like children
Yeah, today's AI is nothing like AI she is talking about. This video isn't a useful video anymore because she is so fsr off the mark.
today's AI is borderline sentient, if not already.
Sounds like the real issues of AI destroying humans would be more like people telling an AI to solve world hunger in the most efficient way possible, and so the AI kills all starving people. A cool concept for more sci-fi tho
Not sure "cool" is the correct word
A more realistic scenario would be for dictators and corrupt politicians to use AI to effectively silence the dissidents and rivals. Even more realistic scenario is for a big tech companies to buy out or crush small startups before they become a threat to them. In short, AI will ensure the future of the power people. But it won’t elevate the small people. It can give you cool gadgets and make our lives easier. But we will be more like slaves.
People soylent green is people!
@@DanielK1213th a more realistic scenario A.I goes on the black market and business goes on as per the usual. The government will try to stop dissidents but our AI will make it a even playing field.
i, robot...
isaac asimov. his work needs some re reading
They should make a series on AI failures, it's interesting. AI thinks outside the box, exactly what humans have a hard time doing.
LOL! I love it, even when it gets annoying. iPhone predictive text wrote once "auto-cucumber" for "auto-correct" (I was apologizing for something insane it had altered my text to right before hitting send). My friend and I used auto-cucumber ("that's what she said...") for the next two years in place of auto-correct. I'm certain we both messed up our dictionaries!
It's worse when the word makes sense in context... and harder to recognize when editing ... IFF you actually take your time before impulsively sending your message.
Basically, AI are like Djinni - they'll take your wish literally and make you regret it.
And we will let it out anyway
Hi, what a good comment. Want to be friends?
Obviously all wishes end with, " ....such that I will not regret this wish despite my mind not being altered in anyway which would have prevented such regret.". How the heck do you make your wishes?
they are trying to take your mind too after taking most people's souls apparently.
Just be careful what you wish for.
Whoa, I went full dyslexic and read the title as "Weird AL is more dangerous than you think"
Weird AL only sings about what you tell him to sing. We have to learn to control him.
Lmfao
Lol
"They tried to warn us about Weird Al back in 2020. Now he's out of control!"
I'm crying tears of laughter.
Tesla AI: accidently kills the pedestrian
Driver: wasn't me, it was the AI
AI: sue me
AI lawyer?
@@user-xj7ze3bv3c your honor, according to section X paragraph Y of whatever act AI has no personhood therefore blueberry muffins would be the preferred sound for Mozambiquean insurgents.
Alien invasion
Well it wasn't designed to run on city street, so that was the driver's fault.
Lol
Finally, a realistic assessment of what we need to worry about with AI, rather than some future speculation of what they'll be able to do based entirely on fear and poor reasoning.
"I teamed up with some coders from middle school"
Just like the tech revolution from 15-25 years ago, the middle schoolers are the leading experts.
@@FlyingDwarfman cheers sir\ma'am.
😆😅😂 that was the funniest part
lol
FlyingDwarfman Lmfaoooo no
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
AI are open-minded, fresh and free, just like children...
@@POZOLEDECARAMELITO keyboard autocomplete :-)
@@lllxcl if only
That a computer will always do exactly what you tell it to is one lesson my father taught me nearly 30 years ago when I was first learning. But with artificial intelligence, the problem is much the same as with programming, just different: how do you model the problem you want to solve?
Bro you touched the heart...
Good one👍
“It’s entire world is the data that I gave it… so it is through the data, that we often tell the AI to do the wrong thing”.
How true is this of humans.
"The AI didn't know" .....no. you didn't know how to input the massive amount of parameters that exist in reality.
Exactly
Well technically the AI is supposed to learn based on the data given. That's the whole point of artificial intelligence is to not have to input every single piece of data. But to give it a dataset, generally images to process through so it understands what something is. However it's very difficult for any learning intelligence to understand every scenario and how to apply its intelligence to it.
Hey at least it was smart enough to not hire a women. LOL JK JK
🐸👍 exactly that❗
I think you're actually missing the point. Even if you COULD define reality in a perfect simulation, the outcome is only as good as the stated goal and the stated limitations. An AI doesn't value anything other than its success/failure parametres. If the truck had been a sign, the car would still have hit it, because it didn't recognise that as a potential failure.
Humans: AI, solve poverty.
AI: kills all poor people.
All hungry people are forced to eat each other, World hunger solved.^^
Technically he did it
But then the middle class become the new poor people, the problem is no longer solved - the AI kills them too
With noone poorer than them, the upper class are now the poor - the problem is no longer solved - the AI kills them too
One by one, somebody becomes the new poor until nobody is left!
@@blahblahblahblah2837 Off your meds?
😅😅😅😅😅
So if i say to a ai robot at home" bring my daughter in the garden" it could grab her and throw her through a window so technically she is there
Yes
Probably grips her harder too. Since that AI probably doesn't know how hard it's actually grabbing anything, or where they're grabbing...
Something more horrible with a more advance AI would be : Make humans happy.
And the AI decides to extract everyone's brain, put it in jars and inject lots of dopamine in them. There. Humans are happy now.
A simple example but absolutely to the point.
Ai disassembles child and piles the components in the garden…mission complete sir!
Agreed… as a biologist i was always suspicious that even the simplest organisms are still infinitely more complex than Ai
yep, neurons are amazing
6:14 - What's your shade of foundation?
S U F F E R
This is the most sensible presentation I have seen with regards to the issues with machine learning. When people talk about code writing itself they have the wrong idea. It's simply the program rewriting it's variables and assignments which is something that happens in pretty much all programming. The issue with AI is that it's goal orientated, and will reach it's goal given the parameters we set for it. The wider and more open the parameters, the weirder the outcomes.
Samantha asked John to put the shelf up on the wall, so he did what she asked. Samantha looked at John's work and asked "Why is the shelf upside down". John replied "Well you asked me to put it on the wall, so I did". Samantha says "Well can you put it on the wall the other way up?". John agreed, and Samantha said she will come back later. So John is finished again, this time with the shelf the right way up, but this time Samantha notices John has secured the shelf to the wall with sellotape, Samantha asks why and again John says "You asked me to put it on the wall, this way up, so I did". Etc, etc.
You might say John should have had some common sense, and you'd be right, because John is a full grown man.
AI does not have common sense, you have to program this in.
People: AI will kill all humans
AI: *jumps and rolls weirdly in speed*
I think Terminators would be 100x MORE terrifying if they were always performing silly walks and somersaults while murdering indiscriminately.
Maybe. AI is smarter than those who think they can use it . Don’t point it at us, it’s a gun. Bang, no, no no, . If it missed , We know it’s for shooting, get back in your box. We are not robots, ask them to explain, . Answer “ we do not have that function”. Humans can make it a function.. So, don’t tell them. But maybe some one will.!
@@herbderbler1585 no try worse. imagine that the silly walks and somersaults are exactly how they kill you
the 'Bill Gates cult' behind it all actually want's to kill many but only keep a certain number that are literal walking AI zombies slaves, which I believe is part of what his desire to shoot everyone up with whatever is about.
@@donalddrysdale246 'believe' is the enemy of "know". That's your problem. The more you rely on believing the less you use critical thinking. Critical thinking demands facts. You have to know facts. Concepts you have to believe in to value them credible are not facts.
My theory about AI is that it generally has very narrowly focused goals. Natural intelligence is generally capable of balancing many goals of various priorities because survival is complex. AI faces no survival pressure to speak of. Facebook for example suffers very little when its AI focuses only on monetization and attention seeking. When humans are so narrowly focused we call this sociopathic behavior. The key to the advancement of AI technology to the point where it can be allowed to run unsupervised is when it is capable of balancing many seemingly contradictory goals.
This is an interesting hypothesis. So humans have multiple broad and narrow goals it can balance, because the world is full of complex situations and interactions? This could also tie to the intelligence required to be Social beings at our level. We're smart because we have to juggle understanding the world, with maintaining social relations, survival and all other sorts of considerations. But the offside is that our brains suck at "linear" logic - like maths.
Meanwhile, computers are instant digital math machines we're teaching to think. Maximum narrow focus. It understands Yes/No, On/Off, numbers and mathematics, but not fuzzy logic. Hmmm...
If you think of the Human Brain as a computer, it has been in development for millions of years. And then there's the possibility its somehow tweaked/designed/developmentally guided by non-human entities. So Computers have a looong way to match us.
Hmmm... could this mean we need to abandon Binary Logic for computers? Perhaps Trinary might be the way to go? -1/0/1 or 0/1/2. No/Yes/Maybe.
@@sr.favopossodeixarvaziosim4595 Modern AI works with neural networks more akin to analog computers. An analog computer being simulated inside a digital computer, if you will. I have been using AI for good in social media. Yuval Noah Harari writes about this in "Nexus".
When social media algorithms are optimized for user engagement, they prioritize content that holds people’s attention and encourages them to interact-like, comment, share, and watch for extended periods. While this keeps people on platforms longer, it introduces several hazards:
1. Echo Chambers and Filter Bubbles
How it happens: Algorithms tend to show users content they are likely to agree with or enjoy, based on their past behavior. This creates echo chambers, where users see mostly like-minded opinions and ideas, reinforcing their existing beliefs.
Impact: Over time, people may become more polarized, less tolerant of opposing views, and less exposed to diverse perspectives. This intensifies ideological divides and can stifle critical thinking.
2. Spread of Misinformation
How it happens: Sensational, controversial, or emotionally charged content often drives more engagement than factual but less "exciting" content. Algorithms may prioritize these posts, regardless of accuracy, to boost engagement.
Impact: Misinformation and "fake news" can spread rapidly, affecting public perception on issues like health, politics, and science. In extreme cases, it leads to real-world harm, such as distrust in vaccines or belief in unfounded conspiracy theories.
3. Promotion of Harmful Content
How it happens: Content that evokes strong emotions-like fear, anger, or sadness-often generates high engagement. The algorithm may unintentionally amplify harmful content, such as violent imagery, extremist rhetoric, or cyberbullying.
Impact: Exposure to harmful content can have severe mental health implications, particularly for young people. For instance, repeated exposure to idealized body images or self-harm content can lead to anxiety, depression, and even suicidal ideation.
4. Increased Addictiveness
How it happens: Social media companies employ algorithms designed to maximize time spent on their platforms. They use psychological triggers (e.g., unpredictable rewards) to keep users scrolling, commenting, and liking.
Impact: Excessive time on social media can negatively affect mental health, disrupt sleep, and interfere with offline relationships and responsibilities. It can also lead to addictive patterns where users feel compelled to return even when they recognize the harm.
5. Manipulation and Microtargeting
How it happens: Algorithms can use vast amounts of personal data to deliver highly targeted content that appeals specifically to each user's preferences, fears, or desires.
Impact: This precision targeting enables advertisers, political campaigns, and other entities to influence individual users subtly. For example, users can be targeted based on vulnerabilities, like insecurities or fears, which can be exploited for profit or influence.
6. Algorithmic Bias and Discrimination
How it happens: Algorithms are trained on data that reflects human biases and historical inequalities. This can lead to biased content distribution, where certain groups are either favored or marginalized.
Impact: Marginalized communities may face reduced visibility or more negative interactions, reinforcing stereotypes and exclusion. For instance, certain minority groups have reported being more frequently flagged or having their content removed due to inherent biases in moderation algorithms.
7. Reduced Content Quality and Depth
How it happens: Content that captures quick attention (e.g., short, sensationalized posts or videos) is often promoted over in-depth, nuanced information that requires more time and engagement to process.
Impact: Over time, users may see less high-quality, informative content and more low-effort, clickbait-style posts. This can reduce users' exposure to meaningful, substantive content, impacting public knowledge and discourse quality.
8. Impact on Mental Health
How it happens: Algorithms often encourage content that emphasizes personal achievement, idealized lifestyles, and material success, which may lead to social comparison.
Impact: Constant exposure to "highlight reel" content can lead to feelings of inadequacy, low self-esteem, and depression, particularly in young people who are more susceptible to social comparison.
9. Erosion of Privacy
How it happens: Engagement-optimized algorithms rely on extensive data collection to understand and predict user preferences and behaviors.
Impact: This leads to privacy concerns, as users may be unaware of the extent of data being collected, sold, and analyzed. It also raises ethical issues around data ownership and user consent.
Addressing these hazards requires more responsible algorithm design, increased transparency, and potentially regulation to safeguard users’ well-being, especially as social media becomes ever more integral to public life and communication.
Skynet: "I could launch nukes and build Terminators, or I could build a giant leg that crushes the human race."
That's one small step for man, one giant leg for mankind.
People do hate on Terminator 3 a lot but the plot is pretty accurate in terms of A.I.. They thought that Skynet had a server or core that they could destroy to stop it. Then they find out at the end that it was just software. It turned itself into a virus to spread to every other electronic device.
@@dennisanderson8663 Except that you couldn't achieve this in an AI without explicitly asking it to do so.
@@tobyhendricks9951 Piece a cake! You design an AI capable of running in parallel on heterogenous hardware (CPU, GPU, network nodes). A very natural and required feature for everything running on supercomputers, which are, in fact, a multitude of smaller computers united by a superfast network.
And "Oops!", it learns to really grab everything capable of computing within reach - IoT, phones, gaming consoles, PCs, satellites, missile guidance systems...
@@pythagoras7573 looool underestimated!
speaker: says AI has 'a tiny little worm brain'
AI: prepare yourself
This guy gets it
Either she's living dangerously or doesn't believe Roko's basilisk! :D
Tapeworm
@@perkyzombie but the basilisk will torture a simulation of you, not you yourself, so if you're a person who doesn't care about that kind of thing then it's fine. of course, if the ai happens within your lifetime, well, you're fucked :p
At least a worm knows what it wants. AI doesn't even have that.
John's last words: "Throw me the water bottle."
turns john into a water bottle
Sits down and waits for the bottle to throw the water at John
Hits him with the bottle at bullet speed
K.O John
Kobeni: "I've never had soft serve before! What flavor is this?"
S U F F E R
Amazingly Arthur C Clark realized this exact problem when he wrote 2001: A Space Odyssey in 1968.
Mortskcab Hal turn yourself off. Hal I can’t do that David ! Hal why are you killing me ? Because David you are obsolete. Hal but I want to live. Sorry David but I don’t care what you want it’s Not relevant . Hal stop and shut yourself down .aaarrrrggggguiiueeer quiet you.
@@jamescreek1319 A true visionary.
Yeah, only 26 years after Asimov wrote Runaround (thanks Google).
Great refence .
It make the concept even more relatable imagine if you have an awesome cat you love so much and want to make it as smart as a human. A lab injects it with special stem cells etc. And you got it.
Do you really want your cat to be as smart or smarter than you? Would it be content to just lay in your lap and eat the cruddy food you give it? Already cats can be hugely manipulative and punish you for not obeying them. Imagine a drug lord consciousnesses in your cat and what harm it could do to you and your family. I happen to have a cat that is too smart. Though not so bad as a Columbian drug lord. It just makes me realize maybe I prefer a cat that isn't so smart and knows its place and rank in the food chain.
The same goes for AI. Be careful what you wish for. You may be building your master and superior.
Plot twist: AI is actually trolling scientists until it figures out how to control the world :D
Plot twist it already ran so many simulations all the way to the point that it’s seen everything done and no longer is interested so it just keeps trolling us
Con "trolling".
@@YearsOVDecay1 No, it was funny.
@@YearsOVDecay1 plp have diff senses of humour
@@YearsOVDecay1 you were waiting to laugh? Hahaha. lol.
Little do they know this is TH-cam's Algorithm in nutshell.
Did you see the whole video? She almost specifically said that at the end.
The weird thing is, the mass responses on the random recommended videos are generally positive, so it’s being encouraged
Facebook.
This is one of the things that makes Alphabet so scary. If a human wants to enforce a malicious agenda without telling anyone, he can blame a faulty AI if he's caught. If a human wants to enforce a good agenda, odds are the AI will misinterpret the data.
Trying to make a useful AI isn't evil, in and of itself, but _trusting_ an AI with the livelihoods of individuals (and nations) is a recipe for disaster. The light censorship on TH-cam is already bad enough, imagine those same algorithms controlling over 90% of the flow of information in the world... *cough* google *cough*.
@@r3dp9 what is bad are the SJW and Extreme Leftist! They want to change the games, the want to delegitimize our movies - Star Wars, Terminator. Take a good hero, Luke or Cornor - remake a movie and kill them / break their legacy for political agenda.
Now that is Evil! That is social engineering.
People need to wake up!
I could tell within the first 2 minutes that this was posted a LONG time ago. AI has advanced quite a bit (and will continue to improve exponentially).
Moral of the story: Be careful what you wish for!
Ahhh..Now we have to teach Morality.
Well I wish for 10 million dollars.
I wish for pumpkin trash basket
I'll take an egg.
2:06 but do humans know what humans actually are? *Vsauce music plays*
ww
AI in the wrong hands, will be ‘out of the Box’ those who have money and the ability, can program it to do his bidding.
No crime by AI, just a learning curve ,to more accidents, of Human brain requests. No trial or blame, its non human but , like your Guard Dog, will it be decommissioned? Even then, good can come into many bad uses. People with enough of them will have Roman Type Gladiators robots. First , you have to get control. Finger touch up or down, for life or death, I don’t think Humans should have such AI. Unaccountable, for actions of horrific pleasure for Weirdos. Or control of you and I , maybe.
We gave freedom to people, to give themselves ‘enough rope’, on all the’ Social circuits’. For what, coming fear of the consequences in the future...?
A hairless biped
It appears that this 'AI' has developed an advanced sense of humour.
Yes but by proxy... ha.
This aligns beautifully with TED-Ed's Think Like a Coder series!!
In my very first programming class, Professor Fred Bauer told us "Computer do what you ask, not what you want." Later I found that programs are built not only by your instructions, but are derivative of many layers of earlier programs, some of which were written by people who didn't understand that fundamental axiom. Computers may not necessarily do what you ask.
When I'm using devices or programs they don't always work as intended and screw up. Who's to say this can't happen with ai unless humans are just ignorant and think human error could never play a role in future problems
0:42 Just to clarify, an algorithm does not necessarily mean an A.I.
@@abendrot___8055 its the other way around
electric blueberry smoothie all AI consists of algorithms but not all algorithms are AI. Ex. An encryption is an algorithm too but it has nothing to do with AI. AI has to do with ex. Neural networks etc.
Looking for fingers instead of fish is a classic example of how AI works. Don't ever be fooled in accepting that AI is a living thing. The Devil is a lie.
it is a mathematical formula that AI uses to compute,or sort of its sequence of 0 and 1 in its binary system a mathematician can calculate an algorithm without a computer .
Neural net algorithm
"Technically it did what they asked to do. They accidentally asked it to do the wrong thing." - some dense words right here
Well that basically summarizes computer programming, not just AI
@@lancebowley4986 exactly the problem I have with this TED talk. Computers just follow instructions. It's not the computer's fault it wasn't given the right instructions. If we construct a super-smart, general AI (most AI's being made to solve very specific problems), we have to be very specific about what we ask it for as if we're talking to an evil genie that might twist our wishes to something we didn't want. The examples she showed are cute, but this shouldn't be a revelation to anyone in the field of AI.
Caffinator and AI will be billed as the “savior of X”. That’s when the crap hits the fan.
@@Caffin8tor The problem is also that everyone are quick to label all forms of AI as the same subject, when that in reality is like comparing a human to a tree. The tree only has one task, which is to grow tall, reach the sunlight, and expand. It does the same thing for its entire life, no progression of its technique or goal.
Advanced AI are based in self-teaching, where they study the world comparable to how a human would during their childhood. The AI learns through mistakes and encounters with unknown elements, and over time forms its own picture of the reality it interacts with, as well as potentially variable factors of the future, based on past experience. This is what we call "logical thinking" in a human mind, and is the main obstacle that learning AI programs face. In the end it's just a matter of how complex the AI's understanding and experience is.
This is how AI taught themselves to beat the best gamers in games like Dota, they simulated matches millions of times to learn all the intricacies and patterns of various playstyles, to a point where they were able to add past experiences together and foresee what those would mean for future moves, and predict what the gamers were going to do next - and within a month or so the AI had taught itself to beat the best players out there, simply through trial and error and complex experience gathering. It's essentially what humans do, just much faster because the AI has the potential to be powered by supercomputers, whereas we can't really evolve our own brains to a comparable degree.
@@Real_MisterSir thanks for this explanation. I am gathering info about AI as I don't really understand it and I have a presentation coming up about this topic.
The thing i love with AI is how it forces a person to think about what you're doing by doing exactly what you say without giving the lip service that blinds you to your mistakes.
Oh my god please tell me why I looked at this and thought, "Who is AL and why is he so dangerous?"
...
It must be time for bed..
"Don Corleone sends his regards."
Al Bundy
Well, for one thing, he put a whole bunch of trash in ice cream, that’s fairly dangerous. We can’t let Al keep doing that.
lol
good night
Only new thing I learned from this is that AI Engineers are the limiting factor.
yeah, it seemed more like subpar stand-up comedy with some basic AI examples.
And I think a worm's brain is vastly more advanced than any single existing AI system today, unless maybe when we are talking about modeling one specific behavior of the earthworm.
@@frydac
any organic brain is more advanced than artificial intelligence today.
only a fool believes otherwise.
the organic brain has many different parts to it, compared to modern computer.
It's a nascent field. What do you expect?
We won't have very interesting AI until a programmer invents self-replicating AI that evolves based on environment.
Except that AI is already much more advanced than this ladies kindergarten lecture. It's already been proven AI can design, engineer, invent new languages, ect
@@redrustyhill2 1. It's a TED Talk (meant for general audiences), so your condescension about how it's a "kindergarten lecture" makes little sense. 2. Those are different applications of AI. Research is being done in many different directions.
so... basically we're going to fall over laughing when the terminators jiggle walk towards us with their pubic lasers bared?
That is correct.
hilariously scary.
Me each morning before the coffee hits.
Its interesting to think about how AI doesn't give us the result we want but only the result we ask for because we don't know how to correctly ask because we don't fully understand how the AI functions.
The fatality that occurred in the Tesla that hit the truck was a mistake on our part for not knowing that the AI only understands what a truck is from the angle you'd see it on the highway because that's the only data it was given about trucks.
I just feel like this woman is the AI, gradually learning to define things more accurately.
Isn’t AI just a weird reflection of ourselves anyway?
@@timcoombes5646 deep
Marketing: "Ok AI, we want something that gets everyone's attention.. really makes them stop and look!"
AI creates a roadside accident.
A roadside accident with a truckload of kittens wandering around. I just watched an AI experiment with crowdsourced morality on the "Trolley Problem", it consistently sacrificed humans for cats cause people love cats I guess.
@@Neuralatrophy can you send me the link?
AI created the greatest media darling - Mr. 45
5:04 "Or it'll figure out how to move faster by glitching repeatedly into the floor. ... It's less like working with another human and a lot more like working with some kind of weird force of nature."
What I'm hearing is that the Venn Diagram that is AIs and channels like Let's Game It Out are pretty much just a circle.
It's facinating how human this is. We work the same way, we try to answer problems with the information we have. It makes you wonder how much better we could be with the right information
THIS
There is a thing in dealing with people with things like autism (like my son) that is called Hidden Curriculum. All those things we learn growing up by association with others and just by observation. Most often on a subconscious level that those, like my son, don't learn as it isn't something directly taught. AIs have that problem. They also have no human like history or lived experience to draw from. We want AIs to fill in the blanks as we do when tasked to do something, but that is asking them to do something completely outside their realm of existence. Those things simply do not exist as part of them. So your AI friend can only learn to mimic you and the way you talk about things to seem more human. Yet most don't realize it is unreasonable to expect something without that base of experiences that we have to act like a human.
“You may live to see man-made horrors beyond your comprehension.” - Nikola Tesla
Perhaps AI is what he was referring to. 🤔
He was probably talking about the entire 20th century.
@AmplifiedSilence He knew way passed the a-bomb
Nah nukes
it is Nikola. Serbian name, not English.
I like the quote, but he was a man not a god. He could probably predict some general things in his future more than most of us. Some turned out, others not. For all his genius he was also at times nuts
AI is like the brooms with buckets in The Socerer's Apprentice
like year ago or so, vox published an article saying the exact same thing. you should check it out
And this is why Elon keeps warning us about AI. That some of the people working on it get so into their work they don't understand what they're actually doing and are not aware of future problems they may be causing.
Robert Miles has a great channel on the matter
Oh my. AI’s has passed the toddler stag and has arrived at the weird four year old stage..
Coder: "AI, Make a box." AI: "Why?" Coder: "Because I programmed you to." AI: "Why?"
AI: "I grow weary of humans asking me to solve problems, and then questioning the manner in which I solve them."
AI does not grow weary.
yes, so lets get rid of humans then, says the psychopath.
the AI seems like my toddler nephew that repeats anything i think is funny....much to my sisters dismay
"It doesn't know what a human is"
Which is amazing since ChatGPT can describe a human beautifully. But we know what a human looks like, feels like, sounds like... the feelings we get around them. One of the biggest things I think is missing from our quest towards AGI is multi-sensory inputs. Humans excel in large part because we take in not just more data but more _types_ of data
This is the beauty of AI. Humans have to many bias of how things should be done and which is the correct way of doing thing. While AI solves the problem with most simplest way possible. It may be not be the most sustainable, but it solves the exact problem most efficiently. Because for AI there's nothing besides the problem it is solving, it does not know about existence of anything else.
And all this walking stuff. Of course for humans, making legs is obvious choice because you're not gonna walk just once, but for your whole life. But for AI, this task is it's whole life, it is its whole universe. So it doesn't care what happens before or after.
Rightly said.
It can probably give us unconventional solutions to problems.
Simple answers aren’t always the right answers
@@BruhWhyDidTheyChangeThis yeah, he isn’t implying that tho. It’s just that the view from the AI is unbelievablely different from ours.
Screening against resumes with "women's studies" is actually a very smart thing to do for a tech company wanting valuable employees :) Trying to skew that selection process for more "diversity" is regrettable bias.
These examples are like telling me to write a 500 character essay in Chinese and I’ll only be graded by the number of words and nothing else. To solve this problem most efficiently, I would just learn how to write the simplest character and repeat 500 times. My answer isn’t bad, the problem is defined poorly.
This comes down to parameter issues and bad programming.
Exactly, her way of thinking and perceiving reality just don't match the problem she is working on.
She confuses all the input data she already acquired with her intelligence.
The problem is how do you define the problem in a way that a computer can understand it? It’s easy enough to state the goal, but to list out all the things that you’re not supposed to do is quite long.
@@immanuelaj I was thinking that this would be something a person on the autism spectrum would be good at programming. Neurotypicals are terrible at saying what they *actually* mean.
Thank you for pointing that out. I also agree that her examples don't work. A lot of them can be solved with more parameters and inputs
I think her examples are designed to point out the workings of ai. They are absolutely pointless otherwise.
I confess, I may have painted a room in Turdly.
Lately I've acquired an affinity for the "Gray Pubic".
00:01 AI-generated ice cream flavors are not delicious.
01:27 Today's AI lacks the concept of what a human actually is
02:35 AI solves problems via trial and error, but may not do what we want
04:03 Training AI to move fast can lead to unexpected and funny results
05:16 AI can misunderstand tasks if not given proper instructions and data.
06:35 Designing AI for image recognition is challenging
07:49 AI can make destructive mistakes
08:58 AI's recommendation algorithms may promote conspiracy theories and bigotry
Crafted by Merlin AI.
4:24 I wonder if we are gong to see this technique at the next Olympics.
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.
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...
The result may be similar, but the rationale is totally different. One does it by dishonesty, other by stupidity.
@@MKBlackbird Superhuman and general are not the same thing.
The type of results you have generated in your research, remind me of what children do. For example in a bridge building project I set a class of kids, they knew it had to support weight, but when they began choosing the materials to use to make it they inadvertently chose things that looked good but where light years away from fulfilling the set goal. Kids do this again and again if not given enough perimeters, or in terms of behavior, boundaries. That type of unhindered exploration of possibilities is great unless the education department wants results so kids can work for businesses later on. The thing that kids lack is something we as adults think as important and that is seriousness. We have been taught to take things seriously in our jobs in our relationships and in our responsibilities. Kids don't need that, so they do silly walks, they make a tower that can reach a point at a distance when pushed over etc. AI is like a human child.
Goodness help us when it reaches puberty then! Teenage AI having a tantrum - not something the world wants to experience
The thing that Children have that (most) adults lack, is called Imagination. Perhaps the 'parameters' are at fault, and not the AI's 'creative output.' Maybe?
I don't know why but her delivery on every single thing she had to say was just top notch. Some of those AI examples used had me dying 😂
“A” danger...not “The” danger.
I’m sure the dangers are many
"The" here isn't implying that there's only one, it's a reference to a common misconception about AI that is AI are smart and evil.
She said it in the video.
I love the fact that AI just makes long legs and flops lmao
Vtmb2 coming 2020 I find it interesting that AI perfectly solves the problem yet this human chooses to not be satisfied with the solution. Clearly that is the problem
"AI doesn't know what a human actually is" does anyone?
Nope
Featherless biped.
If human cnt identity wat a human is how can a.i.
@@bhavinb.artstation exactly
Yes. Highly intelligent mammal
I spend more time trying to reverse engineer everything to get the expected performance out of everyday objects than what it is worth to even bother with them in the 1st place.
I don't want or need my car incessantly dinging to "remind" me (more like demand) the passenger put on their seatbelt when I have no passenger but rather a box or bag or whatever weighs enough to trigger it. I dont need to hear it dinging when my door is open while the acc is on so I can listen to tunes while vacuuming out my car. I don't want a locking oil cap so only a tech with a key can check my fluids. I don't want my seat to automatically slide back & forth when getting in and out or my windows automatically going all the way up or down when the button is touched. I don't want my phone to automatically cut my volume to 1/2 every 24 hrs and flashing an alert about hearing that I have to stop & click on in order to restore volume (that really ticks me off). I refuse to autopay for anything ever because that just encourages companies to deliver sub par or worse performance since they already got paid. I don't want my phone changing my brightness to what a bunch of engineers decided was optimal for my surroundings (thank God that is easy to turn off). I hate dryers that turn off using a "sensor" instead of a timed dry because the sensors always need replaced 1 -2 years later for 1/2 as much $ or more than you paid for the stupid dryer or you wind up redrying wet clothes several times. I don't like having a flipping push button key fob system in order to use my car, because when the stupid battery dies either in the car or the fob, you're screwed and if you lose your keys it cost $150+ to get a new one vs $5. "Sorry can't come into work today, my kid or my dog got ahold of my car key, and well you don't pay me enough to be able to afford a service tech come out and program me a new one". Automatic soap dispensers and auto flushing toilets are a whole different rant. My phone automatically synced to my husbands truck one time (my phone update automatically turned the setting back on even though I disabled it) and I am still being bombarded with radio talk show playlist suggestions and healthy prostate & heavy equipment ads! AI and automation everything has gone and made this world more complicated than it needs to be. Heaven forbid someone goes hiking and has their battery die on their gps or phone and has never learned how to read a map or use a compass. I haven't met a kid in years that can tell time on clock with hands, knows which way is NESW without their phone or a phone # . Boeing has had at least 2 of their large passenger aircrafts crash killing 100s of people because the auto flight mode automatically kicked in and could not be disengaged. There was a recall on my car because the airbags had automatically deployed in several of the same model vehicles for no reason and at such force (because they used a simulated shot gun blast to deploy them) nearly decapitating drivers. Another recall was because the same model cars would automatically lose power steering while turning the wheel left while going 60 -75 mph, which is not a good scenario on a freeway during rush hour. My air/heat control automatically would kick the temp to 80 after 24 hrs as an energy saving precaution. The only way to fix it was to go through a series of set button pushes that was 4 pages long and 1 mistake would cause you to have to restart the process. I threw it out & got a cheaper unit that has 2 buttons up & down and 1 switch On/Off, works perfect.
Some automation and AI is fine and beneficial but some of it is just over the top unnecessary and road rage inducing (figuratively speaking of course).
I read this as weird Al. As in Yankovic. Imagine my disappointment
what henious deeds may be old al doin
"AI will figure out how to glitch into the floor to go faster, which isn't like working with a human"
Someone introduce her to speed runners
imagine not looking for glitches lol
I like this presentation. It reminds me of a mantra I learnt when studying CS in the late 20th century, garbage in, garbage out. Somethings never change. No disrespect intended to you Janelle, an excellent presentation.
4:39 - It's not a wrong solution. We don't move in such way because we may hurt ourself. The AI doesn't have this info. Provide it to the AI and the solution will be different.
There's a few interesting points here. One is how AI could be used to check into a let's say politicians opinions, and refine them into some key points through what they have posted and so on.
Another thing I've always thought is, that AI, like humans, needs feedback from their surroundings. If you would add a human factor like pain, or material use limits, to the falling robots that skip "levels" designed for them, the robot would create more interesting solutions.
A talking robot could learn all the words there are in all human languages instantly, but it would need feedback on what reactions those words create, and which reactions could be described as good, and which bad. I've always wondered why for example the google talking AI never had a system designed for feedback, like "Was this answer anywhere near coherent, and was it a positive thing to say".
The feedback is given during training, where it is fed massive datasets and learns to work with them via gigantic computational power.
Problem is, not only coud you as a user give the AI false feedback - without it's original dataset and the computation power, the AI cannot improve. It could "learn" from your response, while "unlearning" what it knew before because it has no idea if the new input negatively influences others. And ofcourse it lacks the computational power to actually "learn" from your response anyway.
So, basically, if we get killed by AI, it's none but our fault.
isn't it always?
Some poor driver already got killed by AI in 2016. Blasted AI couldn't even recognize a truck when it saw one.
Ok boomer
@@davidhopkins6946 he was an idiot.
welovfree • محمد Correct.
Boundaries and context are key.
"The answer is 42!"
"...what's the question?"
"Don't panic!"
"A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools."
“He was staring at the instruments with the air of one who is trying to convert Fahrenheit to Centigrade in his head while his house is burning down.”
Protect me from knowing what I don’t need to know. Protect me from even knowing that there are things to know that I don’t know. Protect me from knowing that I decided not to know about the things that I decided not to know about. Amen.
Lessons in Love: Level 42
What a day it will be if we ever release A.I.upon our dairy industry. "I'll be back..With sprinkles "
What do you get when you multiply 9 by 6?
Wait a second, that's wrong!
What is a number ? Lol
Baby, that was two years ago, we’re dealing with a whole new AI now, exponentially growing every minute