tv Inside Story Al Jazeera May 28, 2015 11:30pm-12:01am EDT
finals. the sister won in 2009. and look who finished third last year. >> i'm barbara serra, thank you for joining us for the news. ray suarez is next with "inside story". as our machines got smarter and smarter technologyists said hey, don't worry, they on the know what they tell them. when certified smart guys like physicists steven hawking says wait a minute you are playing with fire is it time to be nor careful about artificial intelligence machines smart enough to mistake decisions
without us. one that is we may not like. that's tonight on "inside story". welcome to "inside story". i'm ray suarez our world welcomes new products every day working fast and well in part. they are working without us. machines are learning by what they do today, tomorrow and next week some get queasy digitally enabled smart products will make decisions without being taught and teach other machines without asking for permission. that's artificial intelligence at a time that launched stories for decades. here is al jazeera's jake ward.
>> reporter: every day approximately 3 billion people go to a form of work and bring with them their human abilities - their experience compassion manual dexterity. in a generation how many of these jobs will be done by a machine? in several industries it seems robots could do the work. the jobs are harder than it looks. one of those things is dry cleaning it takes human dexterity to deal with the variety of clothes that we wear. a robot can't look at this know what it is and feed it through a standardized machine that could handle other pieces of clothing. >> as soon as i was born my parents had me on their backs working the machine. by the age of 7 i was putting expungers on hangers, taking ticket at the counter. by the age of 16 i was working on saturday so they could have a day off. my whole life i've pretty much worked at a drycleaners.
>> we need humans to drive us around in taxis. >> they are pushing us to a world where humans are worse than necessary. at the moment we need human resources and judgment and senses letting us know dangerous stuff like when the person in front of us is drunk. let's look at how many are effected. out of a total workforce, 210,000 work in the dry cleaning laundry industry. 233,000 workers, are drivers of taxis. half a million would be in need of new lines of work in just the two industries. >> the u.s. military is a driving force behind robotics and artificial intelligence. a robot can also wield a weapon
and track the troops. >> it's a futuristic agency in charge of insane products. >> the most important was a concept that could divide an attack. they hosted a competition like this. they could drive themselves and they wound up with google's self-driving car. we tend to be afraid of robots much maybe it was the movies the best robot lab in hollywood, and sometimes they can be a little scary. others don't have that. they have an open idea about what a robot could do to help us. the future is exciting. military aspects are alarming. the military built the x 47 b. a pilotless combat plane that can take off and land without
human assistance. and the military held exercises in 2014 experimenting with weaponized robots the thing that hollywood screen writers based their scripts on it would be another generation before artificial generations will make it possible to make a meal or press a shirt and a few years before they drive a car. the irony is killing a human is a simple task and robots can do that now jacob ward joins me from san francisco. welcome back to "inside story". for the purposes of this conversation what do we mean when we talk about intelligence in artificial intelligence. is it autonomy. >> yes. it is autonomy. there has been a tremendous kind of tectonic movement in the last 10, 20 years around robotics and
the notion of art initial intelligence. could we get to the case of singularity, where robots are smarter than we are, and get to a place where they are enhancing our lives without having to anticipate. that idea has sort of fallen out of favour and it became clear how difficult it would be to create a robot. that could learn on its own, that could perceive its surroundings it was just the complexity from the brain. we are getting to a place where data is plentiful. that is the key to teaching on artificial machine. how to receive surroundings and make decisions, and how it gets more and more powerful. we are better at feeding it to robots such that it is learning. is it the ability to make smart
calculations in their mind or is it experience in the sense that humans understand. early on in this age i was speaking to someone who hoped to design ai and he said well building a computer that is like a human brain is hard because we do so many things so quickly because we can call on all our memories, call on the accessions quickly, in the early 1990s, that technologist couldn't imagine the tremendous steps forward in memory that we have taken, in the ability to create big computer brains that can do something like that smell when the difference between a burning pork chop and maybour living room being on fire. >> that's absolutely right. at one point it was unimaginable really that machine could, for instance tell the difference between two natural objects.
an okay and an elm tree. it's easy for a robot to be fed the proportions of a pre-us or -- prius or a manufactured object. perceiving what a tree or a dog is was hard. it's a link of can you teach a rob to the do the short and we do in our minds. when you walk down the street and see a car coming towards you, your brain doesn't look at the headlights and grill and hub caps and say "that's a car." you quickly establish short and and your brain screams car and you jump out of the car, and the mental short cuts make us so applicable intellectually and trying to teach them robots are struggling now. teaching them that the polygons form a corner and that is the corner of a table, and again, the headlights the grill, the hub caps they have to do all of
that. we are trying to teach them to do it in shorthand the way you and i do and we are just a tiny of the fraction of the way there. we have only gotten to the place we can do the simply stuff, but the simple stuff is driving. you and i think it's simple it's teaching robots the shouldn't that you and i use every day. >> making a garment involving competitive motion doing the same thing 500 times a day. will we see those kinds of jobs going away first, because they are easy. >> undoubtedly. the pick and place kind of stuff, the grabbing the thing, putting it in another place, no problem. that is easy. bird-watching. you know being able to track things - that is hard. the irony is that the easy stuff is figuring out there's a person running, and i can shoot at it. that's why all kined of activists are concerned about the military applications
identifying a suspect. taking him out. it's easy for a robot to do. it's the subtler stuff that is hard, that's where we have years and years to do. >> we'll talk about military applications later in the program. jacob ward joins us from san francisco, good to talk to you jake thank you so much for having me. >> you know it's happening. faster and faster processes making it possible for machines to make smarter and smarter decisions independent of human action. the question is where is the line is there a point where you are okay with a thinking machine, and another point where you are not. prospects, promise and perils of asending machines is tonight's "inside story".
companies when drivers broke speed laws even if there has been no accident or mishap and it raises their rates. they learnt how to reliably understand human speech to have realistic spoken exchanges with human being and remain stob ornly tethered unable to intre rate or respond, without a person starting out by giving the instructions setting the boundaries deciding what is possible and what is not. will you have a problem with them since choosing a tart and launching a drone strike. joining us now is bill hib ard, a senior scientist emeritus at the university of wisconsin, madison. when technologists like bill joy or world renowned physicists like steven hawking say to the world yes this is happening, but please be care:
why do they have that twin j of warning. >> well i think there's a variety of problems that arise with artificial intelligence. the one we discussed earlier was our desire to build machines to make our work easier it's a natural desire a consequence of that is there's less work for some people to do. there's other problems that might arise. for instance if you think of a theory you have in your iphone, you can talk to it but you can't talk with it the way you talk to another person. at some point you will be able to talk to a little voice you carry around with you, with some jury. and a lot of people will speak with that voice all the time. that companion will get to know them very well.
the single company or a few companies are providing the companions to the population and the large computer server will be able to know a large number of people very well and that will give that company, whoever it is a lot of power to manipulate society. >> i knew that we would go quickly to the dystopia idea a creepy idea of what the possibilities are. hearing about how they can interact in a genuine way, i visited a lap in the university of southern california where they were trying to figure out how to make companions for shut-out people. they were making tremendous strides in getting home-bound people who might not otherwise have human companionship to keep the subtleness of their muscles,
range of motion in a way that we'd have to send out 1,000 workers to get it done. this was doing it reliability and cheaply. >> yes, it's two sides of the same coin. we are building it to do things that we want. we want to have these companions they can do so much for us. in my own work the ability to search the internet. we can search for it and get it quickly. it enables the search companies to get to know me. they can make an accurate profile of exactly who i am. >> i'm sorry, is part of the problem that technology outpaced the evolution of best practices that go along with it and legal
batteries that would go along with it. >> i listened to an astute lawyer. he said the answer to whether an innovation would be legal depends on whether it is popular. everyone will want these things they'll be useful. i'm 67 years old. i can't wait for a self-driving car, so when i'm too old to drive, i don't have to drive. just the fact that this technology is going to be sow intwined in our lives is what is going to make us vulnerable to that technology. >> how do we take care how do we get the best use out of these wonderful innovations without letting them cross the line? >> i think the most important thing to do is what you are doing, and that is looking into this issue and bringing it out to the public in general. i think the most important thing is for people to have a really clear understanding of what the
technology will be like so that the people at large can be aware of the problems and can and can decide through the political process to what extent it needs to be regulated or kept an eye on. >> people are, friendships, they have misgivings about machines mowing down soldiers. they did in the 1970s, haven't we been here before? >> yes you can think of an analogy with work. machines in the 19th century replace a lot of physical work and mental work. now as you discussed, machines will be doing more mundane mental work and people will be able to retreat the more sophisticated work. there'll come a point, at least i believe there'll be a point
where machines will do any work we do. it's a qualitatively different machine. the same goes with the military situation. when machines can - right now the worry about military weapons is they'll not be able to distinguish friend from foe. there'll be a time in the future when machines will name that distinguishment more accurately and that is the real danger. >> bill of the university of wisconsin, maddison. thank you for joining us. >> thank you so much robots replaced thousands of workers from machines that can learn to fold to machines that make decisions. they are poised to replace thousands more. what it means to you in a minute. this is "inside story".
welcome back to "inside story" on al jazeera america. artificial intelligence this time on the programme. the first time i saw a robot working on an autoassembly line was in the 1980s, there were certain parts of a job putting a car together they could do quickly and decisively. today the plant would have a lot more robots doing more things and few people. there were plenty of jobs we were told that would never be replaced by a smart machine, because there was parts of decision making too subtle for the yes, no stop go binary world. maybe not. martin joins me author of "rise of the robots", as you heard, bill hibbert in the last segment talked about how we first
thought mundane tasks would be replaced by machines now higher order ones are. should we be worried. >> yes there has been a lot of hype given to the issue of superintelligent machines doing things like taking over. we hear elon musk and steven hawking worrying about that. that lies in the future. the impact on jobs is happening now, it doesn't require sign for example technology just an extension on what is happening. most workers in the economy do relatively routine predictable things which is suited to robots and algorithms and technology. we'll see a big impact over the course of a couple of decades there. >> are we in a position where we have to take the good with the bad? i mean we like the fact that a blood monitor can test our sugar
and but just the right amount of insulin into our mix so we remain upright and alert. we like that. we don't have to say machine, manage my blood sugar, it does it without us thinking about it. >> that's right. it's a dual ench sword. -- dual-edge sword. this has made us more wealthy over time. that's why today we are better off than 100 years ago. it's the advance of technology technology is reaching an inflection point, a new age where machines are going to start substituting for loss of workers. on the surface that is not necessarily a bad thing, it may mean we have to do less work or less of the work that we find unpleasant. it creates a distributional issue, a problem where more and
more people can't find a place in the economy, where they don't have skills to find a marketable job and may not have access to an income. >> earlier in our history that was portrayed to workers as something that would make us freer from toil and wealthier and give us leisure time. it turns out not to have done that at all. why not? >> it turns out that in the economy as it exists people aren't saishiated. they want more and more. in order to survive in the economy, you have to work full-time. there are, of course people that work part time for what we find is that very often the people have to string together two or three jobs because we have seen a massive increase in the cost of living. realistically it doesn't appear
that capitalism is going to lead to a world where we can work a few hours a week and everything will be fine. >> exactly. >> exactly. >> earlier in the programme, we heard something about mindful, and managing this technological shift. can it be managed. can it be done in a way that is less destructive. that means the world is not shivering under our feet this way? >> well i think it should be possible to adapt to the economic implications of it. what i advocated and say in my book is in the long run we'll have to move towards a guaranteed income scheme where everyone will have access at least to a minimum income whether they work or not. in the context of today's politics it's radical and seems impossible. the paradox is it's probably inevitable if we do the
adaptation. eventually people will come to see that and there may be a political movement in the direction. >> what are some things that are coming that we haven't thought about. >> we see it across the board. it impacts every aspect. flipping hamburgers sitting in an office driving a car, truck, they are all impacted. one of the interesting things that we see is that the conventional wisdom is the lowest skilled jobs. in fact what we see a many white-collared knowledge jobs are easier to operate. it doesn't require arms or so forth. we are seeing some of the better jobs hit harder. that creates a paradox going
forward. >> writing software that writes software. >> it's researched now. where they do the brains primarily and not the hands, and are susceptible to this. on the other hand people that fold laundry, we'd like to automate that. it is difficult with the technology it would be a long time before we build a robot to do that. >> author of "rise of the robots." i'll have a final thought in a moment of being careful what you wish for. talk back to your television follow us at ajinside-story. twitter, facebook and an intelligent machine will present your comments there. we'd love to hear from you.
once. if you had your job replaced by an intelligent machine, it's probably not resulted as we were told it would in world fairs or disneyland displays or release from human druggingry and increased leisure time. analysing insurance claims filing inscriptions and driving across town it involves some of the discontent that a company repeated activities but you do have a wonder about how the benefits are distributed, and you have to wonder and if you question the pace and breadth. for now, thanks to the people that made the machines that bring me to you tonight, and thanks to the humans that make them work. thanks to you for joining us for "inside story" see you next time i'm ray suarez.
suarez. thailand warns of an alarming level of asylum seekers as it hosts a regional commence on the migrant crisis in south-east asia hello, welcome to al jazeera live from doha. program elizabeth puranam, also ahead... >> it will fall to me it to take responsibility a defiant sepp blatter addresses the court bombs e