I recently finished reading Rise of the Robots: Technology and the Threat of a Jobless Future by Martin Ford. In the book he paints a rather bleak look at the not-too-distant future where robots begin taking over a lot of the jobs that support us. He predicts that within a few decades over 50% of the jobs that are here today will be gone through being done better and cheaper by automation.
Unlike past predictions, he’s not just talking about factory jobs, although those are well on the way to being replaced, too. I read that Apple is working hard to automate as much of the production of the iPhone as possible. But Ford says that the first jobs to disappear will be white collar jobs and that jobs that require physical skills and people-to-people interactions like plumbers, electricians, and auto repair might be the most immune from automation.
It’s a bit glib to use the word robots, because white collar jobs are going to be taken over by the combination of supercomputers and big data processing. And it’s also a bit of a misnomer to say that jobs will be replaced. Some jobs will be 100% supplanted, but for many jobs computers will be able to take over some significant portion of the work currently done. And in doing so companies will need fewer people.
The way to understand how this might happen is to consider the recent announcement that a deep learning computer learned to play master-level chess in 72 hours. There have been computers beating the best chess players in the world for a few decades now, but this is very different. The old computers were good at chess by using brute force and the computer programs looked ahead at millions of possible outcomes before making a move. But the new deep-learning computer learned to play chess the same way that humans do.
And if the computer can learn to play chess it can learn how to do a whole lot of jobs that people do. I have a friend, Danny, who operates a large CPA firm and he has already started his business down this path. He has written programs that have fully automated the process of reconciling bank statements and using that data to produce a set of books and a draft tax return, all with almost no human intervention. This has allowed him to save on labor and probably puts him five years ahead of his competition. But the industry will catch up to him and that is going to eliminate the bookkeepers and accounting clerks who have been doing that function everywhere.
It’s not hard to picture this same computer process eliminating the jobs of almost anybody that manipulates data. If you work in front of a computer screen you probably can and will be replaced. Some of the major news outlets already are using computers to generate sports stories for box scores. These programs can take a box score and turn it into a short word blurb that is hard to distinguish from one written by a human. But the new deep-learning approach is going to go far beyond this very simple task. I just read about an AI program that is helping scientists by searching through millions of published papers to find research related to what they are working on.
Ford is an industry insider and it’s hard to find fault with his conclusions. There are many others making the same prediction, but Ford makes the case more clearly than most others I’ve read. Where I disagree with Ford is in his suggestion of a solution. It’s obvious that if half of the people don’t have work that we will have to find a social solution. He suggest what he calls a Minimum Annual Income given by the government to those who can’t find work. It’s hard to picture with today’s politics that the county is going to step up and pay people to live who don’t work and will probably never work. If not, we face a very bleak future of haves and have-nots.
What I find most dismaying is that as a society we are paying zero attention to this issue while it is right in front of us. We are going to start seeing the jobs disappearing in the coming decade, with the decade following seeing the full brunt of technology replacing people. Take the example of pharmacists, which is typical of the kind of knowledge job that can easily be automated. It’s fairly unlikely that somebody just starting pharmacy school today is going to find lifetime employment in that field. There are already hospitals trialing robot pharmacists and reporting that they operate flawlessly without errors. While all pharmacists won’t go, one would think that the big drugstore chains will replace most pharmacists over the next decade or so. This is just one example and the same thing is going to happen to another thousand job descriptions. As good jobs start disappearing this is going to create possibly the biggest shift that society has ever faced over such a short period.
I don’t get paid anything to write this blog so I don’t worry about becoming unemployed as a writer. But the predictions are that a lot of technical writing will be done by computer soon. I don’t know if a computer is ever going to be as opinionated as me, but that is probably something that can be built in as well. So perhaps one day a computer will take over this blog. I’ll be sure to check in from time to time, though, just to see how I’m doing.