Transcripts For CSPAN Technology Policy Institute- AI Automa

And unemployment is at an alltime low, so there is nothing to worry about. And probably the truth is somewhere between those, inclusive maybe. And so that is what this panel is talking about. And whether itai is a substitute or complement to human labor. I will introduce our guests weekly. Diane bailey is an associate professor at the university of coverswhere she engineering product design, economic development, and she conducts large scale empirical studies. Essen is from the Boston University school of law. His latest book is learning by. Ng ece kamar is a researcher at Microsoft Corporation and focuses on Artificial Intelligence. Rian is a chief economist at google. Been involved in many aspects of the Company Including Corporate Strategy and is also an emeritus professor at the university of california berkeley in three departments, business, economics, and information management. A good place to start is to talk about what we actually mean by Artificial Intelligence. Ag brought this up in the preliminary discussions that this would be good place to start so where are we today with ai . First of all, thank you for is aing me to this panel person who is studying ai and is an involved in policy discussions, im glad to be here to hear the discussion. Artificial intelligence, is a field that is around 17 years old and started in england. Then the father of ai came together and wrote a paper that wed, this is what ai is a going to be able to solve this ai problem in three months . And they had a meeting where they discussed the different applications of ai, him and several applications were discussed. And they realized the problem was going to be a bit more difficult than they thought was going to be. Some things turned out to be easier, andb some things turned out to be harder. And then they realized ok i think we are Getting Better at this. Oh no, i think this is really hard. Im kind of going through this back and forth. So what is ai . There are multiple definitions of ai. When definition i like very much talks about, its the activity of making machines intelligent. And what we mean by intelligence is, for whatever domain you are designing for, we expect the machine to act appropriately within this environment, by testing its environment and acting on a. Thats what we mean by intelligent machines. However, intelligence is a term that we think of humans for. A lot of the abilities humans ai system toct an have. So another definition that is more human focus is having the ability that is very specific to humans. I want to go into why we think ai is so much in the press. When i was a student 10 years ago the advice that was given me was, if you are entering the job market, dont say you are united because ai was a term that seemed like it was poisonous. However that was a time that google was booming and google is in ai company. So what is really going on is , there were two things in the last five to 10 years that came together to make things happening in the field very successful. Are, we have a lot of data now coming from sensors coming from crowdsourcing, coming from human negativities. We have more competition than human activities. We have more competition than ever before. Through this we are now seeing great advances in perception like facial recognition, image recognition. See these paths that are associated with humans, understanding what the objects we say that machines are really getting intelligent. Before going to the labor has been it, ai around, people have been working on for years, and comes and goes in popularity. The last case that time it was popular, someone might have had a panel like this , saying that this is the time is going to work . Or is it really Something Different this time . I think thats the question we are here for to discuss. Nobody knows the answer to that. Around, when you look there are a lot of applications of ai people use every day. All of the optimization of algorithms, but he does not like ai wasnt there. I think ai was good for some werent goodmans at, like understanding probabilities, making trending decisions, making optimization with this and now technique, ai is moving into tasks that, i wasnt good at. Is creating a perception about what other tasks machines may be able to do now. And second, what does this mean for the human jobs . And that is something we need to say. However there is another discussion which is, does this mean we are getting to general Artificial Intelligence . Much customization, i am going to be able to stop all ai tasks . I dont think were getting there, thats my opinion. Jim, you been writing about labor and ai. Tell us your views. We have actually had ai in the workplace, in the marketplace, since 1987, when ai frauds were used to do detection and creditcard systems. Computerve had automation, which isnt so different from ai, since the 1950s. Whats interesting is that we are seeing an acceleration, within the scope of things that the handle, and maybe in the pace of how it is addressing it. So its in the automation. And we perennial have i think, a basic, many people have a basic misunderstanding about what automation means for jobs. Its commonly assumed, some tasks in a job are automated, that jobs are lost in that occupation. And thats simply not true. We can look up manufacturing and we are very well aware that a lot of manufacturing jobs have been lost to automation. In the 1940s there were nearly half a million Cotton Textile workers in the u. S. And now there are only 16,000. Most of that difference is from automation. Some is from global trade, but that is clearly having a dramatic affect on many communities and many workers and their families. But the thing to remember is, automation can also increase half ad we have only got million textile workers because for the previous 100 years automation was accompanied by job growth. So this is strange. How can automation sometimes create more jobs, and sometimes eliminate jobs, and was going on what is that me . And it comes down to demand. At Something Like textile automation at the beginning of the 19th century, the average person had one set of clothing. Automation meant the price of cloth went down, which meant people could afford more, and they bought more, infected but so much more that even though they needed fewer workers to produce the clock, many more workers to produce the cloth, many more workers were employed because there was much more demand for the clock. Him to an a for the price that and a for the price decline would not produce more demand th. Clo if you look at today and what is happening with computer automation, we see lots of examples of how computer automation, just like textile automation, is creating jobs. One of my favorite examples is the bank teller. There was a great untapped demand for getting cash at and the atmions, machine kim along and people assumed that was going to wipe out the bank teller. In fact, we have more bank tellers since the atms were installed. And the reason is, it made it cheaper to operate a bank branch, thanks could operate more branches and serve more people. There was a market demand for that and they built so many more branches that, even though they needed fewer tellers per branch, they were employing many more. Pattern i think is the we are going to continue to see, and many sectors of the economy, not manufacturing, over the next 10 or 20 years. And i think with the Immediate Response is going to be to ai, as well. You,ane, before we go to accelerating changes in a will generate more jobs more quickly. But that is counter to what some others say. One of the differences that the rate of change isnt allowing industries to catch up, allowing the labor market to catch up. Then you are saying the opposite, right . That the faster the change, the better will be. There are two things, yes and no. If you have elastic demand and you are making faster change, productivity improvements are bringing job growth, if you make faster productivity improvements you are going to have faster job growth. It is going to be disruptive, though, in another sense. Maybe overly optimistic the in the way i described things. When the question is, are we going to see mass unemployment in the next 10 or 20 years, the answer is no. Are we going to see lots of individual jobs destroyed . Yes. We are going to see lots of jobs destroyed and others created. Those textile workers in North Carolina need to have skills, need to find jobs, and there are jobs opening up in the rest of the economy. Arelike everywhere else, we saying and will continue to see jobs eliminated. So the acceleration will put more stress on our ability to transition people, to retrain them, to relocate them. I know you are much less optimistic. Let me just push a little bit by asking this. A book called the job talked about,de we were losing a lot of manufacturing jobs and we were talking about how these people need to be retrained into other jobs. For example, we will train them to be bakers and work in a bakery. Is onlyturned out there so much bread and pastries that we can eat and there werent enough for these other jobs for these people to take. And then the language of job retraining started to change from teaching people knew Technical Skills to enter different jobs, to focusing on what they called soft skills. Told,ople started to be the reason you dont have a job is because your Communications Skills arent very good, or this type of thing, you dont work well on a team. As started to put the onus on workers for their lack of self skills, rather than on recognizing that we had a structural shift in the economy and what kinds of jobs were available. Sits asked to weeks ago to on an Engineering Panel to talk about engineering workforce and how they need to be more adaptive to survive in this new economy that we are going to be seeing. A large public institution, the university of texas. So getting an invite to sit on a National Academy of Engineering Panel is a big deal. It means i can put it on my cv, and maybe get a 3 raise instead of a 2 raise. Have some skin in the game. But i turned it down. And i turned it down because i told them i didnt believe in the premise of the panel. Ok . Putting theanel was onus on engineers to become more adaptable, learn more skills, become a quick learner. There telling all of us these things rather than saying, we are going to start seeing some fundamental shifts in the economy and made we start planning for that. All of us. Not just us individuals, running around suddenly becoming more adaptive, better, quick learners, moving up the scale, because there is only so much room at the top. If we think about ai might be true, there is only going to be so many jobs up there and not all of us are equipped to go up that conceptual letter to take at the top youre doing worse worries me, what is going to be left for everyone. There are a lot of problems with job training, and i think could see more complex than that. We got issues by geographic location. Because where you see a lot of the jobs of hearing is. You also have great difficulty, my book is called learning by doing and one of the themes is that a lot of new technologyrelated skills have to be learned on the job. So its not a matter of the classroom, entirely. We have to come up with new ways of getting people experience. But he think we do see plenty of sectors where there are midskill jobs emerging in numbers. Nursingance, nursing, jobs have been in great demand for a long time. Yet it is still very difficult for people to transition into those, and maybe we dont have enough. I think we dont understand what is involved in making these transitions. You are going to bring in the longerterm process. Aboutant to say a word the jobtraining issue which i think is interesting. If the demand for the job this year the skills are here, there are two ways to solve the problem. You can bring the skills up to the demands he can bring the job down. Going on is a lot of through technology because it used to become if you were a catcher you had to know how to make change. They used to be a few were a taxi driver he had to know how to navigate around town. Necessary anymore. It used to be if you were a veterinarian you had to be able to identify 150 breeds of docks. And you can do that with a i come are with your phone for that matter. So this is a big deal because it allows for the kind of onthejob training you are talking about. You drive around town, you learn your way. You learn to make change because thats what the machine tells you. And theres a lot of delivery mechanisms now which are extremely efficient in the onthejob delivery of education. Look at youtube. There are 500 million video views per day of how to videos on youtube. And all that she almost everyone in this audience has picked something in their house has fixed something in her house by going and looking at that youtube video. And these arent solving quadratic equations, their important manuallabor skills the people can learn, how to weld, how to replace a screen door, how to hang a window. If i were going to play devils advocate i might point repairt that is a lot of people that didnt get called in to do work. Hal when you look at the next part of my talk we talk about what happens to them. I want to talk about the theme that released to this discussion. We talk about the demand for a. I. Will reduce the demand for a. I. Will reduce the demand for labor. On the other hand, if you look at the supply of labor, we can find a different story because there is only one social science that can predict 10 years ahead, and that is demography. Everything kind of pales besides that. 1946, that is when the baby boom started. Basically 1946 to 1964. After the baby boom, there was a baby bust. Then there was the echo of the baby boom. You can look through this whole series of population changes and add 65 years to it and we see what is happening now. Namely all those baby boomers who are retiring, followed by the baby bust. What does that mean . Right now the labor force is growing at half the rate of the population. In the decade of the you will 20 20s, see the lowest growth in the labor force since world war ii. When they started measuring it. When you look at the labor force if you restrict immigration, it , is actually going to decline. All those baby boomers are retiring. They expect to continue consuming. Right . But you need some workers somewhere to be producing this , stuff they need to consume. So you have this race going on between automation, which is increasing productivity, and you have the supply of labor which is very, very low to decline. And we have a good in the united dates. Go look at china, japan, korea, germany, italy. They are seeing outright declines in the labor force. It is very, very worrisome from the point of view of the future of their economies. Look at robots. What countries of the most investment in robots . China, japan, korea, germany, italy. They have to have those robots. They have to have some improved efficiency and productivity to produce the stuff their population will be demanding. So that is true as a worldwide phenomenon. By all accounts, unless there is a really big surprises on the automation side, you will see this in the next 25 years. A tight labor market in developed countries, for the next 25 to 30 years. And that is just reading it off the demographics. Scott how far down the line you see that . The labor force becoming more constrained . Hal when you look at the figures around 2060 bc the labor , force growing at the same rate as the population. It is interesting to think this is all because of this huge shock of world war ii. It created this gigantic demographic event that does not work itself out for 100 years. Scott it is the baby boomers fault . Hal of course. [laughter] scott so it seems like so far there are force leopard issues. There are four separate issues. First, a general shortterm versus longterm. Is, is there anything you can do for people who might be im not sure what the right word is displaced in the short run . Does jobtraining play a role when we learned about the effectiveness of that . Then the distribution affects, both shortterm and longterm. And overtime, the demographics and demand for labor, which will swamp everything. So, diane, o sorry, and the inequality issue. Whether it benefits just accrue to a small group. Diane, you turned down this position at the academy. What should their project have focused on to address your concerns . Diane i think what we ought to be paying attention to is power dynamics. You hear a lot if you read about books on a. I. And predictions about jobs, look at the bureau of labor Statistics Data that describes jobs. And it describes the tasks and jobs. Based on that description we will tell you some percentage of jobs are going to be automated or replaced by a. I. Within some period of time. Right . So they are doing that simply based on the description of what people do. No job is just what you do. Every job takes place in an environment that is surrounded by, for example, all kinds of occupational norms and perhaps regulations. I spent a decade studying how engineers are using new computational techniques and software. Things like finite element analysis. Are doing that simply based on the description of what people do. No job is just what you do. Becad how it was changing design and analysis in that field and what it was doing to the workforce. I will give you two quick examples that point out issues of power and when workers have power and how they control Technology Choices made for them and when they are not in power. The people who chose power, it is power held by the government somewhat. If you look at Civil Engineers to design Building Structures design Building Structures like the one we are in their solutions are governed , by strict laws and design regs that involves things like here review and countyreview of plans for building. Because of this building were to because if this building were to fall down, because in this building were to fall down, those of us who survived, would sue. The person responsible is the Senior Engineer who put his professional stamp on the drawings for this building. And because that person faces professional liability, they are very careful about using automation in their work because they know Boston University<\/a> school of law. His latest book is learning by. Ng ece kamar is a researcher at Microsoft Corporation<\/a> and focuses on Artificial Intelligence<\/a>. Rian is a chief economist at google. Been involved in many aspects of the Company Including<\/a> Corporate Strategy<\/a> and is also an emeritus professor at the university of california berkeley in three departments, business, economics, and information management. A good place to start is to talk about what we actually mean by Artificial Intelligence<\/a>. Ag brought this up in the preliminary discussions that this would be good place to start so where are we today with ai . First of all, thank you for is aing me to this panel person who is studying ai and is an involved in policy discussions, im glad to be here to hear the discussion. Artificial intelligence, is a field that is around 17 years old and started in england. Then the father of ai came together and wrote a paper that wed, this is what ai is a going to be able to solve this ai problem in three months . And they had a meeting where they discussed the different applications of ai, him and several applications were discussed. And they realized the problem was going to be a bit more difficult than they thought was going to be. Some things turned out to be easier, andb some things turned out to be harder. And then they realized ok i think we are Getting Better<\/a> at this. Oh no, i think this is really hard. Im kind of going through this back and forth. So what is ai . There are multiple definitions of ai. When definition i like very much talks about, its the activity of making machines intelligent. And what we mean by intelligence is, for whatever domain you are designing for, we expect the machine to act appropriately within this environment, by testing its environment and acting on a. Thats what we mean by intelligent machines. However, intelligence is a term that we think of humans for. A lot of the abilities humans ai system toct an have. So another definition that is more human focus is having the ability that is very specific to humans. I want to go into why we think ai is so much in the press. When i was a student 10 years ago the advice that was given me was, if you are entering the job market, dont say you are united because ai was a term that seemed like it was poisonous. However that was a time that google was booming and google is in ai company. So what is really going on is , there were two things in the last five to 10 years that came together to make things happening in the field very successful. Are, we have a lot of data now coming from sensors coming from crowdsourcing, coming from human negativities. We have more competition than human activities. We have more competition than ever before. Through this we are now seeing great advances in perception like facial recognition, image recognition. See these paths that are associated with humans, understanding what the objects we say that machines are really getting intelligent. Before going to the labor has been it, ai around, people have been working on for years, and comes and goes in popularity. The last case that time it was popular, someone might have had a panel like this , saying that this is the time is going to work . Or is it really Something Different<\/a> this time . I think thats the question we are here for to discuss. Nobody knows the answer to that. Around, when you look there are a lot of applications of ai people use every day. All of the optimization of algorithms, but he does not like ai wasnt there. I think ai was good for some werent goodmans at, like understanding probabilities, making trending decisions, making optimization with this and now technique, ai is moving into tasks that, i wasnt good at. Is creating a perception about what other tasks machines may be able to do now. And second, what does this mean for the human jobs . And that is something we need to say. However there is another discussion which is, does this mean we are getting to general Artificial Intelligence<\/a> . Much customization, i am going to be able to stop all ai tasks . I dont think were getting there, thats my opinion. Jim, you been writing about labor and ai. Tell us your views. We have actually had ai in the workplace, in the marketplace, since 1987, when ai frauds were used to do detection and creditcard systems. Computerve had automation, which isnt so different from ai, since the 1950s. Whats interesting is that we are seeing an acceleration, within the scope of things that the handle, and maybe in the pace of how it is addressing it. So its in the automation. And we perennial have i think, a basic, many people have a basic misunderstanding about what automation means for jobs. Its commonly assumed, some tasks in a job are automated, that jobs are lost in that occupation. And thats simply not true. We can look up manufacturing and we are very well aware that a lot of manufacturing jobs have been lost to automation. In the 1940s there were nearly half a million Cotton Textile<\/a> workers in the u. S. And now there are only 16,000. Most of that difference is from automation. Some is from global trade, but that is clearly having a dramatic affect on many communities and many workers and their families. But the thing to remember is, automation can also increase half ad we have only got million textile workers because for the previous 100 years automation was accompanied by job growth. So this is strange. How can automation sometimes create more jobs, and sometimes eliminate jobs, and was going on what is that me . And it comes down to demand. At Something Like<\/a> textile automation at the beginning of the 19th century, the average person had one set of clothing. Automation meant the price of cloth went down, which meant people could afford more, and they bought more, infected but so much more that even though they needed fewer workers to produce the clock, many more workers to produce the cloth, many more workers were employed because there was much more demand for the clock. Him to an a for the price that and a for the price decline would not produce more demand th. Clo if you look at today and what is happening with computer automation, we see lots of examples of how computer automation, just like textile automation, is creating jobs. One of my favorite examples is the bank teller. There was a great untapped demand for getting cash at and the atmions, machine kim along and people assumed that was going to wipe out the bank teller. In fact, we have more bank tellers since the atms were installed. And the reason is, it made it cheaper to operate a bank branch, thanks could operate more branches and serve more people. There was a market demand for that and they built so many more branches that, even though they needed fewer tellers per branch, they were employing many more. Pattern i think is the we are going to continue to see, and many sectors of the economy, not manufacturing, over the next 10 or 20 years. And i think with the Immediate Response<\/a> is going to be to ai, as well. You,ane, before we go to accelerating changes in a will generate more jobs more quickly. But that is counter to what some others say. One of the differences that the rate of change isnt allowing industries to catch up, allowing the labor market to catch up. Then you are saying the opposite, right . That the faster the change, the better will be. There are two things, yes and no. If you have elastic demand and you are making faster change, productivity improvements are bringing job growth, if you make faster productivity improvements you are going to have faster job growth. It is going to be disruptive, though, in another sense. Maybe overly optimistic the in the way i described things. When the question is, are we going to see mass unemployment in the next 10 or 20 years, the answer is no. Are we going to see lots of individual jobs destroyed . Yes. We are going to see lots of jobs destroyed and others created. Those textile workers in North Carolina<\/a> need to have skills, need to find jobs, and there are jobs opening up in the rest of the economy. Arelike everywhere else, we saying and will continue to see jobs eliminated. So the acceleration will put more stress on our ability to transition people, to retrain them, to relocate them. I know you are much less optimistic. Let me just push a little bit by asking this. A book called the job talked about,de we were losing a lot of manufacturing jobs and we were talking about how these people need to be retrained into other jobs. For example, we will train them to be bakers and work in a bakery. Is onlyturned out there so much bread and pastries that we can eat and there werent enough for these other jobs for these people to take. And then the language of job retraining started to change from teaching people knew Technical Skills<\/a> to enter different jobs, to focusing on what they called soft skills. Told,ople started to be the reason you dont have a job is because your Communications Skills<\/a> arent very good, or this type of thing, you dont work well on a team. As started to put the onus on workers for their lack of self skills, rather than on recognizing that we had a structural shift in the economy and what kinds of jobs were available. Sits asked to weeks ago to on an Engineering Panel<\/a> to talk about engineering workforce and how they need to be more adaptive to survive in this new economy that we are going to be seeing. A large public institution, the university of texas. So getting an invite to sit on a National Academy<\/a> of Engineering Panel<\/a> is a big deal. It means i can put it on my cv, and maybe get a 3 raise instead of a 2 raise. Have some skin in the game. But i turned it down. And i turned it down because i told them i didnt believe in the premise of the panel. Ok . Putting theanel was onus on engineers to become more adaptable, learn more skills, become a quick learner. There telling all of us these things rather than saying, we are going to start seeing some fundamental shifts in the economy and made we start planning for that. All of us. Not just us individuals, running around suddenly becoming more adaptive, better, quick learners, moving up the scale, because there is only so much room at the top. If we think about ai might be true, there is only going to be so many jobs up there and not all of us are equipped to go up that conceptual letter to take at the top youre doing worse worries me, what is going to be left for everyone. There are a lot of problems with job training, and i think could see more complex than that. We got issues by geographic location. Because where you see a lot of the jobs of hearing is. You also have great difficulty, my book is called learning by doing and one of the themes is that a lot of new technologyrelated skills have to be learned on the job. So its not a matter of the classroom, entirely. We have to come up with new ways of getting people experience. But he think we do see plenty of sectors where there are midskill jobs emerging in numbers. Nursingance, nursing, jobs have been in great demand for a long time. Yet it is still very difficult for people to transition into those, and maybe we dont have enough. I think we dont understand what is involved in making these transitions. You are going to bring in the longerterm process. Aboutant to say a word the jobtraining issue which i think is interesting. If the demand for the job this year the skills are here, there are two ways to solve the problem. You can bring the skills up to the demands he can bring the job down. Going on is a lot of through technology because it used to become if you were a catcher you had to know how to make change. They used to be a few were a taxi driver he had to know how to navigate around town. Necessary anymore. It used to be if you were a veterinarian you had to be able to identify 150 breeds of docks. And you can do that with a i come are with your phone for that matter. So this is a big deal because it allows for the kind of onthejob training you are talking about. You drive around town, you learn your way. You learn to make change because thats what the machine tells you. And theres a lot of delivery mechanisms now which are extremely efficient in the onthejob delivery of education. Look at youtube. There are 500 million video views per day of how to videos on youtube. And all that she almost everyone in this audience has picked something in their house has fixed something in her house by going and looking at that youtube video. And these arent solving quadratic equations, their important manuallabor skills the people can learn, how to weld, how to replace a screen door, how to hang a window. If i were going to play devils advocate i might point repairt that is a lot of people that didnt get called in to do work. Hal when you look at the next part of my talk we talk about what happens to them. I want to talk about the theme that released to this discussion. We talk about the demand for a. I. Will reduce the demand for a. I. Will reduce the demand for labor. On the other hand, if you look at the supply of labor, we can find a different story because there is only one social science that can predict 10 years ahead, and that is demography. Everything kind of pales besides that. 1946, that is when the baby boom started. Basically 1946 to 1964. After the baby boom, there was a baby bust. Then there was the echo of the baby boom. You can look through this whole series of population changes and add 65 years to it and we see what is happening now. Namely all those baby boomers who are retiring, followed by the baby bust. What does that mean . Right now the labor force is growing at half the rate of the population. In the decade of the you will 20 20s, see the lowest growth in the labor force since world war ii. When they started measuring it. When you look at the labor force if you restrict immigration, it , is actually going to decline. All those baby boomers are retiring. They expect to continue consuming. Right . But you need some workers somewhere to be producing this , stuff they need to consume. So you have this race going on between automation, which is increasing productivity, and you have the supply of labor which is very, very low to decline. And we have a good in the united dates. Go look at china, japan, korea, germany, italy. They are seeing outright declines in the labor force. It is very, very worrisome from the point of view of the future of their economies. Look at robots. What countries of the most investment in robots . China, japan, korea, germany, italy. They have to have those robots. They have to have some improved efficiency and productivity to produce the stuff their population will be demanding. So that is true as a worldwide phenomenon. By all accounts, unless there is a really big surprises on the automation side, you will see this in the next 25 years. A tight labor market in developed countries, for the next 25 to 30 years. And that is just reading it off the demographics. Scott how far down the line you see that . The labor force becoming more constrained . Hal when you look at the figures around 2060 bc the labor , force growing at the same rate as the population. It is interesting to think this is all because of this huge shock of world war ii. It created this gigantic demographic event that does not work itself out for 100 years. Scott it is the baby boomers fault . Hal of course. [laughter] scott so it seems like so far there are force leopard issues. There are four separate issues. First, a general shortterm versus longterm. Is, is there anything you can do for people who might be im not sure what the right word is displaced in the short run . Does jobtraining play a role when we learned about the effectiveness of that . Then the distribution affects, both shortterm and longterm. And overtime, the demographics and demand for labor, which will swamp everything. So, diane, o sorry, and the inequality issue. Whether it benefits just accrue to a small group. Diane, you turned down this position at the academy. What should their project have focused on to address your concerns . Diane i think what we ought to be paying attention to is power dynamics. You hear a lot if you read about books on a. I. And predictions about jobs, look at the bureau of labor Statistics Data<\/a> that describes jobs. And it describes the tasks and jobs. Based on that description we will tell you some percentage of jobs are going to be automated or replaced by a. I. Within some period of time. Right . So they are doing that simply based on the description of what people do. No job is just what you do. Every job takes place in an environment that is surrounded by, for example, all kinds of occupational norms and perhaps regulations. I spent a decade studying how engineers are using new computational techniques and software. Things like finite element analysis. Are doing that simply based on the description of what people do. No job is just what you do. Becad how it was changing design and analysis in that field and what it was doing to the workforce. I will give you two quick examples that point out issues of power and when workers have power and how they control Technology Choices<\/a> made for them and when they are not in power. The people who chose power, it is power held by the government somewhat. If you look at Civil Engineers<\/a> to design Building Structures<\/a> design Building Structures<\/a> like the one we are in their solutions are governed , by strict laws and design regs that involves things like here review and countyreview of plans for building. Because of this building were to because if this building were to fall down, because in this building were to fall down, those of us who survived, would sue. The person responsible is the Senior Engineer<\/a> who put his professional stamp on the drawings for this building. And because that person faces professional liability, they are very careful about using automation in their work because they know Computer Software<\/a> programs can yield Unrealistic Solutions<\/a> based on unfeasible assumptions. And for Civil Engineers<\/a>, everything is about those assumptions, of how a load travels through buildings, for example. So they have rejected a lot of automation. It is not that they do not use techniques like that, they do but they dont use any , automation between the steps of Engineering Design<\/a> and analysis that we see in other fields. Now i will go to automotive engineers. Scott are you portraying that as a positive thing about engineering . Does that necessarily make us safer or less safe . That they are rejecting these technologies . Diane i think it makes as much safer. I would hesitate to write in an ride in an elevator that a computer had designed and all , the Civil Engineers<\/a> i watched would tell you the same thing. Ok . And it is because you have to watch i can give you countless examples. I do not study things at the macro level i study things at , the micro level. I spend sitting at the elbow of hours engineers while they are designing. And while they do so, i talked to them. The computer crashes, we have time for a short interview. I asked why did they do this thing . At the Automotive Firm<\/a> they had a secret laptop with secret software. They were not supposed to use it anymore. They had it locked in a desk. And the boss did not know about it, because it was restricted in a software programs, that they were so does to use. These are the things i asked them, why do you use these technologies . So i think its a good thing that the legislation was there. Civil engineers took safety they did not take it as a harness. Or something that they wanted to shape, they took it as an ethical obligation and something they were proud of. Something that made them different. Automotive engineers do not have the same kind of things guiding their work. Yes, they have the National Transportation<\/a> safety board, but there is no stamp that is to be put on the vehicles. They just have to pass government tests. Their work has been rationalized. Their work has been digitized. Their work has been computerized in ways changing what is ways that are changing what is happening in the workforce. Very quickly i will just say the engineers i studied since about 2004 have a hiring freeze on analysis engineers in the u. S. They only hire analysis engineers, simulation engineers in india. Because they built a big enter ,here, and i spent months there at that center and months also in michigan. At the center they have offshored the work of building the models. The reason they were able to offshore that work was because we had digitized the models. We have digitized and met the much highest mathemati zed come up and it means that people in india can do the work for a fraction of the cost of a u. S. Engineer. My point about adaptable. How can u. S. Engineers who wanted to do that work be adapted in the u. S. When it would happen with the india . They would have to move there. You cant do that job anymore in the u. S. So that was a change that came about because those engineers did not have the same type of power to control what they do. I will conclude with this group, that does have power radiologists. All your medical scans can be sent abroad for reading and they are. They are often sent at night in when your radiologist does not want to be woken up. They send it to india. In india a radiologist will examine it and it comes back to the u. S. Why do we still have radiologist employed in the u. S. . Because the ara, their professional Association Lobby<\/a> for legislation that sent those scans have to be signed up in signed off in the morning by a boardcertified u. S. Trained radiologist. Scott hal, a response to that . Hal i think the radiology not only can be done in india, but now can be done automatically. That has been true, not just recently actually, that true for a decade or two. Because in a lot of cases recognizing the malignant cell , or Something Like<\/a> that, is really pretty straightforward. It can be done by even relatively untrained labor. , or, there are border cases and lots of things where you might want have some of the supervision you were describing, but it can be turned into basically exercising exclusionary power to keep a privileged position. True. S, i think that is i have often said we would have , driverless vehicles, Autonomous Vehicles<\/a> on the right now if it were not for the darn human drivers. Not to mention the pedestrians that are even worse. [laughter] you have a controlled environment like a freeway, and expressway, a controlled environment is really possible to have Autonomous Vehicles<\/a> right now. It was possible to have Autonomous Vehicles<\/a> at least a decade ago in that context. It is dealing with all the exception handling that is the problem in many cases. Scott feel free to jump in. Ece k discussions about a. I. , there is discussion about super intelligence or taking jobs away. But i worry about something more shortterm or not, which is that as ai applications enter the these ai algorithms have their problems too. How are they going to be handling the shortcomings of ai, will the determinants the how we get the value out of this technology . There is a new excitement about a. I. However that excitement comes with the downside. They are hard for people to understand. This relates very much with the comments about, if i know when my algorithm is doing the wrong thing, then i can override it. Techniques,istical but the algorithms are learning from, large amounts of data, it is quite impossible to understand what the algorithms are going to be doing for each case. On top of that the algorithms , get updated pretty often. For example, tesla car. There are algorithms about how they are going to be driving and they are updated about every week. As the driver you were expected to understand. How should i trust my tesla . So there is a big trust problem between the ai algorithm and the user or the controller or the supervisor of the algorithm. And we need to do a lot of work, transparent,layer in the sense that we can actually create a partnership between the human and the machine, so they can Work Together<\/a> or we had so that we ok, it to the case where, know my algorithm was not doing the right thing, i should override. For example, one of the cases where ai algorithms have been used in public space, is for low and sentencing decisions that happen in our legal system. For a decade now, and it is not anything statistical learning , algorithms have been employed in courts for making sentencing and important decisions. Beautiful article which discusses issues by having such a system working with judges, a great thing for all of us to understand, some of the social issues which come with the use of Artificial Intelligence<\/a> in public spaces. Timeu are a judge under pressures and need to make these decisions wouldnt it be easier , to just agree with the machine decision . What do you gain by overwriting it . And theverride it person then commits a crime, you are at fault. You need to really think hard about the balance of opinions control, and be responsible for these decisions, when you actually have a team of ai engineers making the decisions who was responsible in that case . Algorithms can be quite biased. For example they were analyzing , statistical techniques for making decisions for africanamericans and whites in society. Scott sorry to interrupt. These are huge issues. Lets try to stay more narrowly focused. But still to take that seem though howthat theme , in your work do you try to make humans and a. I. Complementary . Whether it is in labor or understanding . Ece we need to move away from the thought that a. I. Will automate and humans will adapt. That is not likely to happen. That is not what needs to happen. I think what we need to work on is, yes, we are going to be working on automation and getting these capabilities for ai, but we also have to think really hard about the design of the middle layer in terms of the aix winning at health to the human, ai explaining itself to the human, and humans having transparency with the machine, and working on the coordination. Thats the only people get to a a right equation. I think there are some other issues too. Think about the exceptions in self driving vehicles. Very often black swan events are hard to predict. You need this huge amount of data to come up with a reliable estimate about how safe the vehicle is. You think about self driving cars. That is guiding it algorithms. But the Insurance Companies<\/a> or government regulars with the public in general dont have access. Were the public in general. It is hard to understand what the actual risk is without some kind of data sharing. There is a new set of issues emerging in terms of data transparency. Hal i think you are exactly right. There will be demand for interoperability, and data sharing for safety reasons. Look at the Airline Industry<\/a> as an example. When it is an airline event there is immediately an investigation by several different parties with different constituencies and interests and they try to resolve the cause and make sure it doesnt happen again. I think that same set of procedures and infrastructure will be carried over into this , context without much objection. I mean who is going to stand up , to object to that kind of investigation . One thing about the news, when you look at the journalism on a. I. , they were picking the they are always picking interesting cases. Poker, all of these games and other things. If you want to get an idea of what is really done in the ordinary sorts of cases, go look le, a company which sets up Machine Learning<\/a> competitions. One Company Might<\/a> say we will offer a prize of 1 million to the party best able to predict hospital readmission within 90 days, using this data set. And i have to say, two qualifications. I was an Angel Investor<\/a> in the company, and it was recently acquired by google. So [laughter] nothing to do with each other those two facts. , it is interesting because you see things like housing appraisal. Zillow put up a data set of housing features and values and put up the best model for pricing houses. Youtube put up 4. 5 million videos and has a contest to predict what people are doing in the videos. Even with still images we have Good Technology<\/a> for that now. We do not have very Good Technology<\/a> for videos, four people moving around, are they exercising . Dancing . Fighting, whatever . The hospital example that i mentioned recognizing leaves, counting plankton. There are just all sorts of applications, there are 230 something applications to get an idea of what is going on in ordinary business practice. It is quite interesting. So i think we will see this diffusing. Reallyre will be the exciting cases like the driverless vehicles and the being batch go champions humans being the go champions. There is a lot of automation going on as well. Ece a. I. Has a longterm problem. This comes to your point we use where you say that there are a lot of cases. In the real world, a. I. Has to do with a lot of edge cases. If you look at the cases were you have a lot of data, the distribution of the data, for some cases we have a lot of data. In the cases where we need edge, there is not a lot of data. The common techniques we use, the statistical learning techniques are good to learn the pattern of distribution and not much detail. So to get it right we really , need to think hard about how to get it right. It could be through the collection of edge cases, creating data sets of those edge cases, sharing them within the industry, but we also need to think about some kind of collection of techniques working together. Not only sticking to one techniques but a collection , working together. And other human supervision to get those cases right. Because just saying, ok we are , getting 95 accuracy on the data set. It does not necessarily mean that the application will be providing value to you in the short term. That comes to your question about why we have not seen productivity affects from a. I. Yet. When you think about those, you need to think about, what is the point where im going to be able to get value from this technology . That is a different question. Scott i want to go to questions. Diane i want to return one thing to bear in mind is to Pay Attention<\/a> to the rhetoric used around a. I. I agree the media focuses on the fascinating cases. But i also think the Tech Companies<\/a> without a lot of rhetoric other on. Self driving vehicles, private i bet all of you can tell me how many Motor Vehicle<\/a> deaths per year we expect to save. It is use in every article, at one point it was 1. 2 million lives. You might not note that as the number of Motor Vehicle<\/a> deaths per year worldwide. The number in the u. S. We dont care about the rest of the world, it is not that, i would just like to explain why they are different. In the u. S. , it is 35,000 deaths are year. To let you know what the number said i was situated between what is right below it and right above it. Right below it, 5 lower, deaths by falling down. Right above it, 25 higher, death by poisoning. Ok . We dont see a. I. Solutions for falling down or being poisoned. We see Artificial Intelligence<\/a> solutions for Motor Vehicle<\/a> deaths. Let us talk about that. Motor vehicle deaths reached a peak in the 1970s and we have been decreasing ever sent. Why . Because mechanical and electrical improvement in that have been ongoing, and also regulations. Dui laws, sobriety checkpoints, mandatory seat will. Seatbtory seat hel elts. You have to ask yourself why are we therefore so interested in self driving vehicles . The thing about sitting lines is not really it. That problem is solving itself. If you look at commercial truck drivers, the first group scott the number of deaths from automobiles is going down, but that does not mean that self driving cars would further reduce that . Diane off driving vehicles might help to reduce the number but the number we are talking , about realistically is 35,000 in this country. Then you can look at other countries. At the university of michigans Transportation Institute<\/a> put out a map of the 25 countries with the most Motor Vehicle<\/a> deaths per year, in if you look at that 2014. Map, you get a sense of what the roads in these countries look like. There is no way that because of things like hal brought up, they are chaotic roads. You are not going to have them. I go to india a lot, i have for the last 20 years. There is no way you will have selfdriving car is in india unless the chairman and ceo of suzuki, which is the largest automaker in india, he says the same thing there was no where we are going to have self driving vehicles in india, they will bring chaos on the roads. We have other ways we can bring these numbers down. I dont think the real motivation of why we are going after selfdriving vehicles is to reduce deaths. I think there is another motivation and i would like to have a conversation. We could really redesign our cities and improve and and some some ofnd and end our Climate Change<\/a> problems. Wouldnt that be wonderful . Lets have that conversation. Hal i will say if you look at the examples you gave, they have this common theme. Advances in monitoring. So people can fall over, be , monitored, and i just saw a pill bottle that warns you about what you are taking and whether you already took this this morning and all sorts of other things. There are a lot of cases i dont , know if i would call it a. I. , but i would say technology. Technological advances in one form or another that help people live safer, more productive lives. In a way, this issue of about selfdriving cars, the problem is dealing with all of these exceptions. I asked the team if you break for squirrels . That is a decision. Avoiding a squirrel could cause lots of damage elsewhere. They say no. Do you break for dogs . They say how big is the dog . [laughter] the break for deer . Of course because there are huge number of accidents involving deer. We need an automated deer to avoid the cars in the will have a safer road. [laughter] scott we have a little time left. Larry . I was a cheerleader in high school. Is it on . All right. Topic,come back to the the demand for labor. Days,n my graduate school a Nobel Prize Winner<\/a> threw out to the class the horse. The horse paradox, or the horse story. He pointed out that between 1900 and 1940, the population of horses in the United States<\/a> have gone way down. Why . Because of the internal combustion engine. We had horseless vehicles rather than personless vehicles for personal transportation outside of urban areas, shortterm haul, hauling freight in urban areas. Cars, trucks, tractors. It did not totally reduce or eliminate all horses, there is still a small population left, but the market clearing wage for horses went way down, below reproduction sustaining levels. Fortunately horses have shorter , lives. Fortunately we feel differently about how to deal with surplus horses. So the question is, are we confident, and hal, i think your point about how labor supply may modify or a familiar rate but ,till, modify or ameliorate but the issue of the demand for labor and arguably we have been so lucky over the past two centuries that Technology Shifts<\/a> have not really diminished the demand for labor. If anything, they have pushed the demand for labor off, but we cannot rule out the possibility of a Technology Change<\/a> doing to , humans what the technology didge between 1900 and 1940 to horses. Scott what happens if we are horses or the population have not decreased . Right. That is the market clearing wage falls way, way, way though. Then what . There are two issues here. Fallsone is, it is difficult to retrain horses. [laughter] james it is not a direct comparison. We can certainly look at occupations that have largely gone away. And we have taken those people and absorb them in other capacities, and that has generally been true. Why has it been true . If you look at individual occupations, it is a story about demand. When demand is elastic there is , job growth. When demand starts getting inelastic, further automation leads to job decline. That can be dramatic. So, in a 10 year or 20 year timeframe, we are not going to see a dramatic change in the nature of demand. But, it is a good question further out. Are we going to basically be able to satiate demand in one market after another so there are no jobs left . And that really raises the broader philosophical question about what it is that humans want. John maynard keynes and leonti thinkthe 1930s, and i leonti talked about by this time we were going to have huge amount of leisure time and technological unemployment. If you look at what is happened to leisure time, yes, the work week has declined fairly dramatically since the 1800s when you had a 72 hour work week. Now we are down to 34 hours. But it has been remarkably slow. And you have to ask, ultimately because asel, it is some level, people are getting value there is demand for things. There is demand for things they will either consume or demand for how they obtain their leisure. And leisure enjoyment. So the technology of leisure has dramatically increased in terms , of video games, movies,. And all these various things the question comes down to in 50 years, is there going to be anything humans want that machines cannot deliver . Think, we are at a philosophical level. Do humans want interaction with other humans . Anthe personal role important aspect of what humans want both in terms of what they , are doing for gainful employment and in terms of how they want to consume . Would deign think i to answer that, i can simply point to some very bright people like john maynard keynes, who had a hard time imagining all the additional things humans might want that might cause them not to have a 10 hour work week. Hal i just wanted to say a couple of problems on that. One is 80 of the u. S. Economy is service. Clearly people want to be , involved with other people. There are lots of services you automate pretty easy. I do not really need a person to lead me to my table at a restaurant. You can flash little arrows on the floor to make it happen. But we want that. People want those services. The work week, nothing is written in stone about a fiveday work week. In mexico, if you look at the oecd countries, in mexico the , work week in mexico is 45 hours. In the u. S. , 37 hours. In the netherlands, 30 hours. France is lightly behind at 33. You can take some of that increased productivity and leisure if you want. Absolutely, no question about it, and he would get very few objections ive inked if people said, hey, let us have every weekend the 83day weekend. Which is pretty much what is technologicallypossible today. That could easily happen if people chose to move in that direction. One last word about the first invasion of the robots. The first invasion was around the 1880s to 1910 or so. It was the mastech robots like washing machines, dryers, dishwashers, lawnmowers, sewing machines, all of those mechanical innovations that made productive. R more until you saw women in particular as she mentioned shifting from the household into , the labor market. And you have seen this tremendous increase in output from a household basis, partly due to that kind of innovation and automation. I have introduced myself as an economic historian. And besides the secret handshake which we have, we are supposed to tell people about how things were the same in the past. I dont think a. I. Is distinct from say the bow and arrow, which is a substitute for spears or the horse, which is a , substitute for messengers on foot. Of course these technologies have implications. The horse was an instrument of the aristocracy because they were expensive to maintain for a while. But i think on the whole these , robots, which is what a shovel robot. Shovel is a size, they are to be viewed with optimism. Here is a number which ought to be in everyones mind. It is not true the number of jobs created or lost in the United States<\/a> is what is reported every month. About 200,000 jobs when things are good. Minus 200,000 jobs when things are bad. In fact, every year, 20 million job or lost in the United States<\/a>. 150 is in a workforce of million. Or 160 million. So there is a tremendous amount of churning. Quite appropriately. In 2000, 130,000 people employed in video stores. So i urge you all to be of good cheer. [laughter] scott probably have time for one more question. That comes back to something ece said at the beginning. Which is whether we are moving towards a general Artificial Intelligence<\/a>, as she called it. It is a generalpurpose technology or a shovel . There is a huge difference. Diane they have argued that if they have argued that if a look act other Technological Innovations<\/a> like the mccormick reaper, it transformed the culture and through a lot of left a lot of people out of work. But because we do not have simultaneous innovations going on in other industries, there were places for them to go pretty quickly. His argument is that what we will see with Artificial Intelligence<\/a> is simultaneous innovations across all industries so there is never left to run. Would you say you dont agree with that idea . No, and i would like to know how [indiscernible] diane how do know they were happening across i think he is looking at the present. [indiscernible] [laughter] its a highly unpredictable thing. No one knew the invention of gunpowder in china would radically change the position of the armored aristocracy. Scott we need to wrap up. So we will take a short break, say 10 minutes and start the next panel. Because if the next panel goes , over we will miss the eclipse and we will not at anyone out to see it. So be back here, right 10 20. Join me in thanking the panel. [applause] [crowd noise] 202 6280184 cspans washington journal. Join the discussion. Live monday, join cspan from the Brookings Institution<\/a> for the results of a Public Opinion<\/a> survey on american and japanese views in north korea. Also a discussion on the military actions concerning north Koreas Nuclear<\/a> program, monday at 10 a. M. Eastern, here on cspan. President trump is at camp david this weekend to meet with Republican Leaders<\/a> and members of his cabinet to discuss the legislative agenda and their priorities for the year ahead. These pictures were shared yesterday. A presse group held conferen w","publisher":{"@type":"Organization","name":"archive.org","logo":{"@type":"ImageObject","width":"800","height":"600","url":"\/\/ia600108.us.archive.org\/23\/items\/CSPAN_20180106_220700_Technology_Policy_Institute-_AI_Automation__Jobs\/CSPAN_20180106_220700_Technology_Policy_Institute-_AI_Automation__Jobs.thumbs\/CSPAN_20180106_220700_Technology_Policy_Institute-_AI_Automation__Jobs_000001.jpg"}},"autauthor":{"@type":"Organization"},"author":{"sameAs":"archive.org","name":"archive.org"}}],"coverageEndTime":"20240630T12:35:10+00:00"}