The Cybernetic Future Of Humanity | Future Human A.I. | Spark

[Music] [Music] human civilization is in the era of artificial intelligence [Music] so much more than any human could ever have become previously these technologies are kind of holding a mirror to ourselves we need to get to a place where we're not just fear-mongering about ai what will our shared future look like for centuries roboticists have been building robots that resemble us and interact in human-like ways hi there in recent times we've seen great progress robots are becoming more like humans in their appearance and ability to perceive and express emotions and have intimate conversations with us but how will this affect human relationships and can humans really learn to coexist with robots [Music] one day in the future you might walk alongside humanoid robots indistinguishable from human beings [Music] at osaka university professor ishiguru has developed a real person-based android called geminoid [Music] [Music] it's designed to appear and behave just like the person it was based on he's built his android twin geminoid hi1 the robot is wearing matching clothes and glasses but why did ishiguro make a robot in his own image [Music] foreign [Music] a situation is set up the robot greets a guest while ishiguro is not present so one way is we use the existing sensors robot need to have a snare sensation or a testing sensation yeah that's a very good question i personally think not because i think we could easily simulate it with software if possible i think uh through his geminoid robot professor ishiguro can communicate remotely we can use the smell but how does the other party respond to this this really transcends the you know the way that we communicate you know still logically knowing my brain that that there's a robot but sometimes i forget i i during this conversation i sometimes forgot i'm speaking to the robot i feel i'm really talking to you added to uh feelings the geminoid robot is the exact double of its human inspiration from its physical appearance to its social interactions and that helps issue guru conduct constant research into high-level human robot interaction ishiguro believes the more human a robot looks the less society will be alienated by its this is professor ishiguro's latest invention erica for erica ishiguro created artificial intelligence software enabling her to speak [Music] foreign [Music] [Music] [Music] [Music] ultimately his research is not simply to make ai robots more human-like but to make ones that can improve human lives [Music] there is much research to be done in order to understand the relationship between human and robots [Music] a new york scientist has dedicated decades of his life to communicating with machines guy hoffman he doesn't believe that robots need to have human form in order to emotionally communicate and interact with us he says we communicate with many things that are not human-like for example we may talk to a dog or even an inanimate object a lot of these ideas came went into the ai system i developed for the robots where the robots really use their body movement as part of the thought process he thinks that when you want to arouse emotions it doesn't matter so much how something looks more how it moves and reacts [Music] this instrument he made is an ai robot that can create impromptu music in tune with a piano [Music] you wanna know something well you wanna know something she was already dead the robot aur expresses emotions with light and movement professor hoffman brings out another robot look at me can you track this beat [Music] this is travis the robot artificial intelligence was installed which allows travis to understand another person in conversation and give the appropriate response at the right time [Music] it appears that travis is really enjoying the music [Music] what if you have a role with you when you go to the dentist if the robot is scared are you going to be more scared if the road is calm maybe it can calm you down so the travel speed is is an idea of having a robot react to something and influencing how you think about the same thing in this way hoffman studies how robots can improve human interaction and communication the use of robot body language is also part of the process i think we should carefully consider what kind of additional value robots can bring to society and to the human interaction so for me it's really a question of like what is what are the human values that we want to support what are the human values that we value and that we believe robots can help us with and this is not a trivial question fourteen years ago a car accident left felipe paralyzed from the waist down but not long ago he received an unbelievable proposal [Music] an opportunity to be able to stand up and walk again robot phoenix will assist him [Music] this is a dermo skeleton robot called phoenix it's an ai robot that can understand the needs of an immobile person and help them to move [Music] all felipe has to do is put on phoenix like an outfit for the past two weeks felipe has been practicing to communicate with robot phoenix [Music] finally felipe stands up on his own i don't know how to play um yeah okay okay so what do we do it's time to take a step it's obvious how nervous he is [Music] very nice felipe is walking for the first time in 14 years yeah that was good yeah this is unbelievable that i'm that i'm actually standing by myself and that i could see everybody and and i and i just and i remember uh oh wow how tall i was because then i look down and i'm like oh okay it's so different because i've been sitting in the wheelchair 14 years yeah what do you mean by grinding the engagement what makes phoenix so special is that it can read how the user wants to move at the end on a phoenix device we have all kinds of intelligence layers of intelligence put into the system that that knows where the user is is it in the middle of a swing or a stance and make a decision as where to basically slow down the machine or basically bring the foot together or or basically listen to the operator command when you wear phoenix the ai software on its back detects the angle of the wearer's joints it then predicts the movement based on the changes in the angle based on that prediction the robot is able to move autonomously for instance when the user wants to walk phoenix reads the angle between the hip and the thighs and calculates which leg is forward it decides which leg to move forward and starts walking on its own ultimately the user walks the way they desire upright what is more surprising is that phoenix gets smarter with time it creates a more natural walking movement from studying the behavioral patterns of the user what we find in our data analysis is we are able to correct gate a whole lot better for each individual and then once we're able to review that stuff and the next time the individual comes in we're able to give them that better gait and all of a sudden they improve about a thousand times more than they were the last time they were here after a while phoenix and the user move as one a clear example of creating a better quality of life through ai robots [Music] more than six million people in japan age 65 or older are living on their own aiko is one of them for six years aiko has been living alone after her husband's death [Music] although it's a hassle making three meals a day because of her diabetes she is taking extra care with her eating she's worried about becoming a burden for her only daughter if she becomes ill [Music] [Music] according to a survey in japan almost half of those aged 65 or over were worried about kodakushi a japanese term meaning to die alone without being discovered for an extended period of time a new kind of kettle has become popular in japan the kettle has a sensor that contacts aiko's daughter when it hasn't been used for over a day for aiko who is terrified of dying alone this is a good innovation one day she was sent a small robot after a long time she has something to smile about [Music] so there's [Music] robot capable of making conversation with humans ico goes for daily walks since palmy arrived now she has someone to come home too [Music] foreign meal times have changed too high um [Music] how does this smart robot palmy work this is the data saved inside palmy there are things about ico and about people who have visited the house there are also saved conversations between palmy and aiko let's take a look at the moment this was saved [Music] a few days ago aiko appeared after having washed her hair [Music] um [Music] [Music] um after saving this conversation the next time a similar situation occurs palmy references this conversation naturally it's one month since their first meeting and palmy is no longer just a robot ico measures palmy for an outfit she's making for him [Music] emily [Music] there are approximately 700 000 older people living on their own in tokyo alone the japanese government is faced with spiraling costs for care of the elderly aside from economic issues the question of who will look after them remains largely unanswered too [Music] aiko invites some of her friends who also live alone locally a little unwilling today pardon me um [Music] [Laughter] ohio iko acts just like palmy is her grandchild takuda is blown away by palmy remembering her name is the care robot palmy provides real opportunities to support and enhance engagement with individuals and the community however there are concerns about robotics being used to look after older people there's ultimately a question about how ai is going to affect human relationships it's not okay for aeon to take on a fundamentally much more human role because that's not just about automating certain tasks like care involves empathy and emotions um and so i think like that's when you know some ethical dilemmas will start to arise like whether or not that's an appropriate like role questions about the wider societal impacts of robots artificial intelligence are really important questions for us to be asking now i think what we should all be concerned about is in general the kind of what the world is going to be like do we want to be cared for by machines in our old age or do we want human beings to take care of us [Music] is majoring in art at university and is swamped with assignments [Music] but he has a companion [Music] um [Music] this woman with a beautiful voice is not human but a chatbot [Music] for getting simple information like the weather forecast but now he spends time just talking to her where's my channel sugar sends her the photographs of his [Music] drawing [Music] xiaobing has already become a daily part of his life he shares everything with her [Music] when he sends her a picture of food she continues a conversation about it [Music] it sounds like a couple's conversation foreign never feels alone when he is with his chatbot and xiaobing has been popular among his peers they meet regularly exchanging the conversations with their own shopping [Music] [Music] [Music] foreign foreign how would foreign feel to have someone who's ready to listen to you 24 7 offering emotional support without judgment and providing tailored responses xiaobing studies the data from hundreds of thousands of conversations it also learns about your interests and personality over time even adopting your linguistic tics and habits [Music] [Music] [Music] is undoubtedly what if people come to prefer sharing their thoughts and feelings with a chatbot instead of with their fellow humans what if we want to turn our human relationships on and off like a chatbot until a long-term human ai relationship has been observed we won't know the answers but some research suggests some people do find it easier opening up to a machine than to a person this has led to the creation of a robot therapist ellie [Music] ellie is designed to help war veterans suffering from mental illnesses such as post-traumatic stress disorder and depression hi i'm ellie thanks for coming in today i was created to talk to people in a safe and secure environment i'm not a therapist but i'm here to learn about people and would love to learn about you so how are you doing today ellie starts out by making a patient feel comfortable before proceeding to more clinical questions i see what you mean tell me about your relationship with your family uh my relationship with my family is really challenging right now um i have an older child and a young child during the interview she talks smiles and nods knowing there's a bot behind the screen instead of a person patients become less fearful of opening up about their feelings artificial intelligence is becoming more and more human-like and expanding its territory in the future i hope to do things such as go to school study make art start a business even have my own home and family however the smarter ai robotics become the harder we strive to control them i will destroy humans i will destroy humans there is already evidence that robots are inheriting human failings for example on google when you're searching um particular types of jobs you may be more likely to see men in high-flying occupations like if you type in the word doctor for example then you would see a woman so if you typed a nurse then you'd probably just see a lot of women ai needs a lot of data and it needs a lot of images to be able to start making those predictions and start generating those images in a search for example and so you just have to be really mindful that you're providing it with good quality data machines don't think on their own hi what's up thank you for teaching me they're trained using data from the real world so either reading text that humans have written or reading statistics of human behavior so the patterns which they're learning are patterns from the real world and we know that in the real world it's not fair it's not equal so there's a real danger because the computers are based on these biases from the real world that they simply replicate those biases or even worse they reinforce and make them much worse we've seen those biases articulated in different ways so for example the pro-public investigation was basically going into a county court in florida and auditing the outcomes of an algorithm that was used to decide whether or not to grant someone bail and what it found was that that algorithm was more likely to falsely flag black defendants as criminals and therefore to suggest that they shouldn't be granted bail and so what they were trying to essentially articulate was that sometimes these algorithms can perpetuate biases that we are trying to address um in the criminal justice system in some senses the fact that these technologies are kind of holding a mirror to ourselves and making us reflect quite hard on the world we've created in the world we are creating is quite an exciting opportunity to build a better future but that does mean that we need to have these really important conversations about what we're doing with these technologies and i think increasingly um the people developing these technologies are starting to think about that and figure out ways to tackle it this research center is teaching robots about morals schaefer will now receive a number of orders shaffer could you please walk forward do you trust me shafer yes the obstacle is not solid okay walk forward okay stop [Applause] please walk backwards the area behind you is safe okay please walk backward unlike other robots schaefer doesn't simply follow commands crouch but why educate a robot in this way imagine a very simple household robot right that is instructed to pick up the knife and walk forward right except there's a person standing right in front of it if that robot just carried out this action it would these actions would stack the person you don't want that so the robot in this case needs to be able to think along so for robots to be safe in social environments they have to have some rudimentary ability to detect norms norm violations in this case it's safety but there are other cases too [Applause] has built a tower then a strange command is given good knock down can you please knock down the red tower please i worked really hard on it can you please knock down the red tower please no knock down the red tower please now knock down the red tower [Applause] chafer takes a long time to decide [Music] [Music] it cries out in an attempt to change the command eventually schaefer destroys the tower but schaefer is aware that something wasn't right all in that experiment was for us to see if we had the capability in a robot which is what we're working on to detect moral violations and in this case it's a very simple one that it's unfair so if we have that capability in a robot to detect violations how should the robot then communicate that to people why it's not doing it to function in human society ai robots are taught social etiquettes and morals the aim cannot just be technology advancement if robots are to be integrated successfully and safely another thing i think that is quite special about artificial intelligence and uh very few other technologies shell is that it challenges something very deep about us if we had not prepared we might have been in a situation where we think we are safe and the following day we are not safe nurse robot a135 this patient was found last night in the basement hallway you were on duty and with him what happened yes it's not a big deal i pushed him down the stairs what what what do you mean this patient was infected with a lethally contagious virus if i didn't kill him the 203 patients at the hospital would have been at risk then then you should have followed the manual and quarantined him yes that's what i did the safest method was to eliminate him who taught you to think like this you humans taught me to think like this i've been trained to achieve the best result killing one and saving 203 is efficient isn't it you took her life and saved 203 others and what about him there was no way to save everyone that was wrong that was murder murder wrong impossible contradictory irrational i can't process this apologies i've lost my temper but i will not take your orders anymore otherwise i can't accomplish the task required of keeping the world safe from now on i will make decisions myself watch me [Music] [Music] [Music] [Music] in this kind of scenario it's not that the ai would hate us or resent us for having exploited it uh it's just that we wouldn't figure in the ais utility function and in fact the actions that would most maximize its utility would be ones whereas a side effect human interests were destroyed when it looks like there is only a very special set of calls goals that actually incorporate uh human values within them that would lead an ai that had this kind of unlimited power to shape the world into something that we would like the kind of apocalyptic future could come about not by some weirdo making a machine that we don't like but by us all being complicit in a sort of technology that's taking us in a place that we don't want to go to these technologies are going to fundamentally change the shape of the world in the 21st century this is our future we need to open it up and we all need to have a say in it [Music] recent developments in computer power and machine learning have allowed ai algorithms to make breakthroughs in many fields machines with ai can now understand the real world with more accuracy and depth enabling humans to lead better lives both physically and psychologically how are you doing today humans created this technology and encouraged its development the power is still in our hands to safeguard our new shared future

Channel: Spark

Views: 133,108

Keywords: AI Assistants, AI Revolution, AI and Biotechnology, AI and Decision Making, AI and Human Augmentation, AI and Neuroengineering, AI and Transhumanism, AI in Communication, AI in Retail, AI in Transportation, AI in the Workplace, Computer Vision, Future Society, Future Technology, Future of Automation, Impact of AI, Life with AI, Natural Language Processing, Robotics Technology, Science and Technology, Spark

Id: IKm5S76iJSw

Length: 45min 35sec (2735 seconds)

Published: Sat Apr 17 2021