Doug:
Many people involved in Artificial Intelligence (AI), Machine Learning (ML), and Robotics (all research areas that I work in) believe that the newest AI tools (sometimes dubbed "Deep Learning") are already having an impact on jobs, and are likely going to have a large impact at all levels of employment (from blue-collar to white-collar) soon. But the Republican administration seems to be down playing this obvious conclusion. Republican Treasury Secretary Steve Mnuchin just said that AI taking jobs from humans is "not even on our radar screen." He continued: "I'm not worried at all" about robots displacing humans in the near future. "In fact I'm optimistic."
Why would he say that? The idea of humans vs. machines is a well-known Hollywood trope, and would seem to be a good match for this administration's motivation by political scare tactics and conspiracy Tweets. So why not stoke this fire? Well, this is a double-edged robotic sword: machine automation saves companies money, while at the same time it eliminates human jobs. And if the political narrative is that the Dealmaker in Chief is keeping jobs in the US only to be performed by US robots, then that is a more complicated story.
Some might suggest that we should begin talking about Universal Basic Income, also known as Guaranteed Minimum Income: the idea that everyone would get paid a enough of a basic salary to live on, independent of what they are doing. This idea may have some merit in the Age of Automation.
David:
Robotics has been having an effect on jobs for a long time. Since before the days of Henry Ford, automation and technology has made for a more productive and efficient workforce. Perhaps Mnuchin is thinking along those lines.
Doug:
It isn't robotics (the hardware). It is the software. The "brains" of the robotics systems are working very well in that they can do things that haven't been able to be done by machines before.
David:
I agree it's the reality that computers are moving into the white-collar market now, and with increasing frequency, and I agree that has more people worried. As government, at all levels, has threatened to make workers more expensive, the idea of replacing them with machines has gained momentum. This is not an idle threat either. Just the possibility that increased wages may be mandated across the board has finally made self-contained kiosks a common feature at airports and fast-food restaurants. These machines have an up-front cost and maintenance costs that had made them less attractive for some companies. Service-industry companies also would prefer to have people deal with their customers for improved satisfaction scores. But as workers become oppressively expensive because of increased health-care mandates, wage mandates, and regulatory mandates, the scales have tipped conclusively in favor of machines. Once those financial outlays for machines are made, and the kiosks and automation are in place, the deal is sealed. Those jobs won't be coming back even if labor becomes cheap and plentiful again. Who does this hurt? Unskilled and poorly educated workers, which, for the most part, equals minorities and the poor.
Doug:
No, I think you are wrong on who this will hurt. There are lots of jobs involving making decisions that will be replaced by machines. I think that includes: buyers, sellers, investors, schedulers, and other currently very well paid jobs.
David:
So we agree that robotics and computers will eliminate both white and blue collar jobs.
Doug:
Absolutely. But the point I make about Guaranteed Minimum Income is that nothing will help this situation. There is no party platform that can save these jobs from automation. So, something will have to change. Either our society will support fewer people, or we will have to spread the wealth to many more people than we do today.
David:
Or new jobs will emerge in a new market. The world is always changing, and jobs that exist today didn't exist 100 years ago. And jobs that were common 100 years ago are long gone today. You usually get what you pay for. Paying people to be unemployed will result in more unemployment. But people are very creative, when they need to be. I'm confident new markets will appear, perhaps in the service industry. The world isn't static.
Doug:
Oh, I'm sure new jobs will appear: robot maintenance, robot psychologists, robot massage therapy, etc. But maybe we can build some robots that run on coal!
David:
Or, a country could just outlaw robotics, or outlaw the replacement of human jobs with machines. I think that's a bad decision, but it could be done, and would eliminate unemployment.
Doug:
No one could make an American company pay workers over buying robots. Wouldn't fly.
David:
Microsoft Founder Bill Gates has proposed making companies pay taxes on their robots to make up for the income tax losses that would come about by all of those formerly tax-paying workers losing their jobs. There wouldn't be legislation to force a company to pay people instead of robots. But there certainly could be legislation to make robots more expensive than workers.
Please note that I'm not supporting this idea, but it's an idea that people like Bill Gates find feasible. His idea actually taxes innovation and productivity. When you tax something, you get less of it.
Doug:
I worked with Microsoft during their dabbling into robotics. Didn't work out. Failed worse than Microsoft Bob. And Gates also thought that Clippy was a good, robotic assistant. Taxes on robots---not a viable idea.
David:
I think at this juncture in our technology, space industry jobs may be just around the corner. In the same way that automobiles created vast new industry jobs, from tires, gas stations, oil changing stations, auto-repairs, electronics, and restoration companies, space travel and exploration will certainly lead to all sorts of new and currently unknown jobs. The future holds great promise.
Doug:
Space as our driving economic future? I don't think so. That will be a playground for the 1%, the super-rich.
David:
Just like automobiles were.....before Henry Ford brought them to the masses with innovation.
Doug:
Some may think that the coming machine learning and robotics revolution is really caused by hype and misunderstanding. I'm not sure we are on the cusp of the so-called Technological Singularity where machines, in an instant, will surpass us all with innovation. But I do believe that it is possible. In any event, I think we need to brace for a very different social structure.
David:
The problem is much bigger, and will affect other countries much more than us. Technology has shown a steady progression for centuries, and our free society has adapted. Technology has made us more productive, and has increased our quality of life. But large portions of the world depend upon manufacturing for the employment of their workers. China and most of Central and South America depend significantly on low-skilled (and cheap) labor for manufacturing. As these jobs become automated, we may see a significant number of unemployed workers (perhaps as much as 50% or more of their populations). This will have significant world-wide effects. Although I see future job markets opening up in new venues, like space and electric vehicles, the short-term losses before that comes about may be huge. World leaders need to be looking at this reality now. Robotic advancements are moving at lightning speed.
Paying people to do nothing is not an answer. Where does the money come from? Why didn't we just pay buggy-whip makers a wage for doing nothing when cars came along? Those people developed new skills for different jobs that developed as the world changed. As our world changes, the challenge will be the same: we need to develop training and education for the new world. Rather than paying people to remain unskilled and unemployed, perhaps we should work harder to make sure students are getting a better education. Perhaps we should require more tech studies in higher education, or switch funding to tech and job training schools.
Once 50% of everyone in of all of the countries South of our border are unemployed, and jobs here are more scarce, we may be glad to have a border wall to make immigration more manageable.
Doug:
Mexico has like 3% unemployment (they are doing better than us):
We have had more Mexican immigrants leave the US than are coming here for a few years. So, I think we are likely to fare worse. But this isn't just about saving buggy-whip makers. This is going to change society. But it is going to take some politicians with imagination. Or, perhaps the politicians will be robots, too. They would probably do a better job. I for one, welcome our new robot overlords.
David:
Hmmm. Robot politicians.
Unbiased, and able to actually evaluate the data and agree to the best solutions, despite lobbyists appeals? Unable to be bought or bribed? But instead of "overlords", let's program them to maximize individual freedom. You might actually be on to something...
Doug:
Seriously though, these systems are not unbiased---they actually codify our biases into software. There are many cases where the training of the system captured the biased views of the makers. For example:
Many people involved in Artificial Intelligence (AI), Machine Learning (ML), and Robotics (all research areas that I work in) believe that the newest AI tools (sometimes dubbed "Deep Learning") are already having an impact on jobs, and are likely going to have a large impact at all levels of employment (from blue-collar to white-collar) soon. But the Republican administration seems to be down playing this obvious conclusion. Republican Treasury Secretary Steve Mnuchin just said that AI taking jobs from humans is "not even on our radar screen." He continued: "I'm not worried at all" about robots displacing humans in the near future. "In fact I'm optimistic."
Why would he say that? The idea of humans vs. machines is a well-known Hollywood trope, and would seem to be a good match for this administration's motivation by political scare tactics and conspiracy Tweets. So why not stoke this fire? Well, this is a double-edged robotic sword: machine automation saves companies money, while at the same time it eliminates human jobs. And if the political narrative is that the Dealmaker in Chief is keeping jobs in the US only to be performed by US robots, then that is a more complicated story.
Some might suggest that we should begin talking about Universal Basic Income, also known as Guaranteed Minimum Income: the idea that everyone would get paid a enough of a basic salary to live on, independent of what they are doing. This idea may have some merit in the Age of Automation.
David:
Robotics has been having an effect on jobs for a long time. Since before the days of Henry Ford, automation and technology has made for a more productive and efficient workforce. Perhaps Mnuchin is thinking along those lines.
Doug:
It isn't robotics (the hardware). It is the software. The "brains" of the robotics systems are working very well in that they can do things that haven't been able to be done by machines before.
David:
I agree it's the reality that computers are moving into the white-collar market now, and with increasing frequency, and I agree that has more people worried. As government, at all levels, has threatened to make workers more expensive, the idea of replacing them with machines has gained momentum. This is not an idle threat either. Just the possibility that increased wages may be mandated across the board has finally made self-contained kiosks a common feature at airports and fast-food restaurants. These machines have an up-front cost and maintenance costs that had made them less attractive for some companies. Service-industry companies also would prefer to have people deal with their customers for improved satisfaction scores. But as workers become oppressively expensive because of increased health-care mandates, wage mandates, and regulatory mandates, the scales have tipped conclusively in favor of machines. Once those financial outlays for machines are made, and the kiosks and automation are in place, the deal is sealed. Those jobs won't be coming back even if labor becomes cheap and plentiful again. Who does this hurt? Unskilled and poorly educated workers, which, for the most part, equals minorities and the poor.
Doug:
No, I think you are wrong on who this will hurt. There are lots of jobs involving making decisions that will be replaced by machines. I think that includes: buyers, sellers, investors, schedulers, and other currently very well paid jobs.
David:
So we agree that robotics and computers will eliminate both white and blue collar jobs.
Doug:
Absolutely. But the point I make about Guaranteed Minimum Income is that nothing will help this situation. There is no party platform that can save these jobs from automation. So, something will have to change. Either our society will support fewer people, or we will have to spread the wealth to many more people than we do today.
David:
Or new jobs will emerge in a new market. The world is always changing, and jobs that exist today didn't exist 100 years ago. And jobs that were common 100 years ago are long gone today. You usually get what you pay for. Paying people to be unemployed will result in more unemployment. But people are very creative, when they need to be. I'm confident new markets will appear, perhaps in the service industry. The world isn't static.
Doug:
Oh, I'm sure new jobs will appear: robot maintenance, robot psychologists, robot massage therapy, etc. But maybe we can build some robots that run on coal!
David:
Or, a country could just outlaw robotics, or outlaw the replacement of human jobs with machines. I think that's a bad decision, but it could be done, and would eliminate unemployment.
Doug:
No one could make an American company pay workers over buying robots. Wouldn't fly.
David:
Microsoft Founder Bill Gates has proposed making companies pay taxes on their robots to make up for the income tax losses that would come about by all of those formerly tax-paying workers losing their jobs. There wouldn't be legislation to force a company to pay people instead of robots. But there certainly could be legislation to make robots more expensive than workers.
Please note that I'm not supporting this idea, but it's an idea that people like Bill Gates find feasible. His idea actually taxes innovation and productivity. When you tax something, you get less of it.
Doug:
I worked with Microsoft during their dabbling into robotics. Didn't work out. Failed worse than Microsoft Bob. And Gates also thought that Clippy was a good, robotic assistant. Taxes on robots---not a viable idea.
David:
I think at this juncture in our technology, space industry jobs may be just around the corner. In the same way that automobiles created vast new industry jobs, from tires, gas stations, oil changing stations, auto-repairs, electronics, and restoration companies, space travel and exploration will certainly lead to all sorts of new and currently unknown jobs. The future holds great promise.
Doug:
Space as our driving economic future? I don't think so. That will be a playground for the 1%, the super-rich.
David:
Just like automobiles were.....before Henry Ford brought them to the masses with innovation.
Doug:
Some may think that the coming machine learning and robotics revolution is really caused by hype and misunderstanding. I'm not sure we are on the cusp of the so-called Technological Singularity where machines, in an instant, will surpass us all with innovation. But I do believe that it is possible. In any event, I think we need to brace for a very different social structure.
David:
The problem is much bigger, and will affect other countries much more than us. Technology has shown a steady progression for centuries, and our free society has adapted. Technology has made us more productive, and has increased our quality of life. But large portions of the world depend upon manufacturing for the employment of their workers. China and most of Central and South America depend significantly on low-skilled (and cheap) labor for manufacturing. As these jobs become automated, we may see a significant number of unemployed workers (perhaps as much as 50% or more of their populations). This will have significant world-wide effects. Although I see future job markets opening up in new venues, like space and electric vehicles, the short-term losses before that comes about may be huge. World leaders need to be looking at this reality now. Robotic advancements are moving at lightning speed.
Paying people to do nothing is not an answer. Where does the money come from? Why didn't we just pay buggy-whip makers a wage for doing nothing when cars came along? Those people developed new skills for different jobs that developed as the world changed. As our world changes, the challenge will be the same: we need to develop training and education for the new world. Rather than paying people to remain unskilled and unemployed, perhaps we should work harder to make sure students are getting a better education. Perhaps we should require more tech studies in higher education, or switch funding to tech and job training schools.
Once 50% of everyone in of all of the countries South of our border are unemployed, and jobs here are more scarce, we may be glad to have a border wall to make immigration more manageable.
Doug:
Mexico has like 3% unemployment (they are doing better than us):
We have had more Mexican immigrants leave the US than are coming here for a few years. So, I think we are likely to fare worse. But this isn't just about saving buggy-whip makers. This is going to change society. But it is going to take some politicians with imagination. Or, perhaps the politicians will be robots, too. They would probably do a better job. I for one, welcome our new robot overlords.
David:
Hmmm. Robot politicians.
Unbiased, and able to actually evaluate the data and agree to the best solutions, despite lobbyists appeals? Unable to be bought or bribed? But instead of "overlords", let's program them to maximize individual freedom. You might actually be on to something...
Doug:
Seriously though, these systems are not unbiased---they actually codify our biases into software. There are many cases where the training of the system captured the biased views of the makers. For example:
- computer algorithms that allegedly tag black defendants as more likely to commit a future crime
- data used to train the software relies too heavily on photos of white people
- tricking Facebook's code into believing fake stories that appealed to right-wing voters during the U.S. election
But perhaps once these learning system become self-adjusting (and self-aware), they can better understand where (and why) they fail.
David:
And then recognize how much more pathetic humans are than they are, and kill us all.
David:
And then recognize how much more pathetic humans are than they are, and kill us all.
No comments:
Post a Comment
Please be kind and respectful. Thanks!