This article's meat is based on some critical assumptions - flawed ones.
Firstly, like most doom-and-gloom technology-obsoletes-humans and technology-steals-jobs articles, the writer assumes all these jobs will be replaced *instantly*. This is clearly wrong, for several reasons.
First, the major corporations that'd be buying the robots are risk-averse. They'll let someone else try - and be burned by - such a scheme before they try it themselves. This might take place over ten or more years.
Secondly, he assumes that this entire block of jobs can be replaced all at once, which is also clearly wrong. They all require varying sophisticated levels of working artificial intelligence. Unfortunately, we cannot assume robots will become capable of handling *all* these jobs at the same time. AI is like nuclear fusion power plants, in ever since the 1950s experts have been saying "it'll be ready in 10 years", and ten years later they're still saying "it'll be ready in 10 years", and so on. It is likely that improvements will continue to be incremental, as they have been so far with industrial robots. Robots capable of taking voice orders from anyone who walks in the door, making your burger, and working the register are the kind of robots that will be perfected *last*.
Third, he assumes that a robot worker will be cheaper than a human worker, and that the rise of robots will not create any jobs to replace those jobs they displace. This is also clearly wrong. Human replacement will take more than a 1-to-1 ratio at first, as the first ones will not be as versatile as humans - they'll be more customized towards doing a specific task. Checkout line robots won't also be pulling shopping carts out of the parking lot and stocking the shelves, you'll need a few custom bots for each job. If the cost of buying and supplying power to a bunch of robots is more than the cost of a minimum-wage human employee, the robots won't get bought. Plus the diversity of robot types would slow the economy of scale of production, keeping the prices up until their widespread adoption.
When robots DO start to become worth buying, they'll need humans to keep them in service - robot repair is a hard enough AI problem that, again, that'd be the *last* type of job robots would be able to replace. As an additional bonus, the human repairmen would probably make a better salary than the minimum wage jobs being lost. There will also, of course, be a spike in the number of robot engineers and robot programmers and robot company advertising firms and robot company markters and salesmen and managers and so on. There will be more business for insurance companies - hey, you want to protect that robot investment! bots make great vandalism targets and it'll probably be illegal for them to defend themselves. There will be more business for lawyers - hey! this robot rolled over my foot, this robot dripped oil in my burger! - as, again, we expect the first models to be imperfect. And as human jobs would be those requiring more skill, there would be more teaching jobs.
Fourth, he forgets that such a massive change in our economic structure would also likely affect the minimum wage. If there are no grunt-work jobs left, then the new jobs would require a level of skill such that the minimum wage would be raised quite a bit - a huge benefit to those human workers with jobs one tier up from those being filled by robots.
Fifth, he doesn't look long term enough. Total automation of all the grunt work would increase the overall efficiency of the system to a level where it would become attractive to shift our economy to a slightly different system altogether. Sort of a hybrid socialist one - hey, if the farms are nearly free to run, might as well give every citizen some free rations of staple foods every month. If construction is nearly free, why have homelessness? Give those who can't afford a house a one-room economy apartment. The economy would still be capitalist at heart - because if you want to improve your situation, you've got to work for it - but the basic costs of living would go way down. In other words, you wouldn't need to earn as much, therefore you wouldn't need to work as much. More people would be working part-time, hence the same quantity of work would employ more people.
hmm. a few important flaws (Score:4, Insightful)
Firstly, like most doom-and-gloom technology-obsoletes-humans and technology-steals-jobs articles, the writer assumes all these jobs will be replaced *instantly*. This is clearly wrong, for several reasons.
First, the major corporations that'd be buying the robots are risk-averse. They'll let someone else try - and be burned by - such a scheme before they try it themselves. This might take place over ten or more years.
Secondly, he assumes that this entire block of jobs can be replaced all at once, which is also clearly wrong. They all require varying sophisticated levels of working artificial intelligence. Unfortunately, we cannot assume robots will become capable of handling *all* these jobs at the same time. AI is like nuclear fusion power plants, in ever since the 1950s experts have been saying "it'll be ready in 10 years", and ten years later they're still saying "it'll be ready in 10 years", and so on. It is likely that improvements will continue to be incremental, as they have been so far with industrial robots. Robots capable of taking voice orders from anyone who walks in the door, making your burger, and working the register are the kind of robots that will be perfected *last*.
Third, he assumes that a robot worker will be cheaper than a human worker, and that the rise of robots will not create any jobs to replace those jobs they displace. This is also clearly wrong. Human replacement will take more than a 1-to-1 ratio at first, as the first ones will not be as versatile as humans - they'll be more customized towards doing a specific task. Checkout line robots won't also be pulling shopping carts out of the parking lot and stocking the shelves, you'll need a few custom bots for each job. If the cost of buying and supplying power to a bunch of robots is more than the cost of a minimum-wage human employee, the robots won't get bought. Plus the diversity of robot types would slow the economy of scale of production, keeping the prices up until their widespread adoption.
When robots DO start to become worth buying, they'll need humans to keep them in service - robot repair is a hard enough AI problem that, again, that'd be the *last* type of job robots would be able to replace. As an additional bonus, the human repairmen would probably make a better salary than the minimum wage jobs being lost. There will also, of course, be a spike in the number of robot engineers and robot programmers and robot company advertising firms and robot company markters and salesmen and managers and so on. There will be more business for insurance companies - hey, you want to protect that robot investment! bots make great vandalism targets and it'll probably be illegal for them to defend themselves. There will be more business for lawyers - hey! this robot rolled over my foot, this robot dripped oil in my burger! - as, again, we expect the first models to be imperfect. And as human jobs would be those requiring more skill, there would be more teaching jobs.
Fourth, he forgets that such a massive change in our economic structure would also likely affect the minimum wage. If there are no grunt-work jobs left, then the new jobs would require a level of skill such that the minimum wage would be raised quite a bit - a huge benefit to those human workers with jobs one tier up from those being filled by robots.
Fifth, he doesn't look long term enough. Total automation of all the grunt work would increase the overall efficiency of the system to a level where it would become attractive to shift our economy to a slightly different system altogether. Sort of a hybrid socialist one - hey, if the farms are nearly free to run, might as well give every citizen some free rations of staple foods every month. If construction is nearly free, why have homelessness? Give those who can't afford a house a one-room economy apartment. The economy would still be capitalist at heart - because if you want to improve your situation, you've got to work for it - but the basic costs of living would go way down. In other words, you wouldn't need to earn as much, therefore you wouldn't need to work as much. More people would be working part-time, hence the same quantity of work would employ more people.