A spectre is haunting America: the spectre of communism. In a world where more and more tasks are being automated, and more and more people are seeing the skills that separate them from the lower rungs of society reduced to a few lines of code on a computer, more and more people are starting to ask: what makes me more valuable to the company than the guy two floors down who makes half as much as I do? So let’s take a look at this issue: what separates a white collar worker from a blue collar worker, how that will change in the future, and what we can do about it as new social and labor classes emerge due to automation.
Conventional wisdom tells us that some people are paid more than others because society values their skills more. This notion is pretty meaningless – it’s easy to talk about society and social structures, but who is really valuing their skills? It’s the guy who writes the paycheck at the end of the day. The only aspect of society that comes into this is the ability to find a guy willing to write that paycheck. The value that society places on our skills is simply the highest value that someone is willing to pay for them, in a vacuum. But then, if someone gets a worse offer for a job than he could’ve gotten in other circumstances, does his value change? Being able to negotiate a salary seems totally irrelevant to the actual work, and yet people accept this as being linked to the skill you’re being paid to employ, at least implicitly. So then, your value is not derived from how much society values you, but rather from how much they are forced to pay for your work, and specifically from how much you are paid.
But this seems rather circular. In the end, we haven’t answered the question of who determines what you’re paid, only that we aren’t paid because of some sort of intrinsic value derived from the type of our work. The answer to how our salaries are chosen is a simple case of opportunity costs. Tech workers are paid a lot because they have the ability to leave their companies to start their own startups, which could be worth many millions of dollars. Let’s say each person who works at Google has a 1% chance of starting a $50 million/year business. Then, it makes sense to pay that person about $200,000 in salary — the employer takes a premium because he’s willing to hedge the employee’s risk of failing on his own, in exchange for the employee producing valuable work. Similarly, banking executives and CEOs have a lot of credibility in their space and have plenty of options, such as transferring to other companies and opening their own firms — their salaries reflect this. Someone who works as a teacher or a coal miner has relatively few alternative options to make more money, even at a reasonable risk, so they are paid less. The corporate world takes as much as it can, and the people who can make the most in this system are the ones who can live without it.
With the advance of automation, however, a new truth is exposed: the system also relies on us. Employers will naturally turn to automation once it becomes cheaper than hiring workers – after all, they pay us as little as they can, so why should they care who or what they’re paying? In the short term, we can easily automate simple jobs like cashiers, librarians, low-level bankers, HR, college admissions officers, and so forth — auto-checkout is simple, a lot of banking work is monotonous, and creating algorithms to decide who should be hired or admitted based on a number of definitive and quantifiable criteria is very easy. Notice that some of these professions are not exactly blue-collar or low-salary jobs, and that’s because the amount of money they’re making is dependent on their number of other employment options, not the difficulty of automating their work. So what happens when the big businesses decide to fire these men and women? Simple, they now have to go it on their own, and take the risks that they eschewed by choosing to be employed earlier. Since even more people are fronting the same risk now, a very small proportion of them will be successful. What happens to the failures? As it stands, they will be unemployed indefinitely and probably have to shift to a different area of work, where they will be paid less. Or, those that only have skills for HR work, for example, will have to go hungry for a while.
In view of this, a natural question to ask about the job hierarchy is: what jobs are the hardest to automate? I suggest that these are the ones which most highly require a working understanding of humans. After all, if we could build a machine that understands human thought, rather than just how to answer a question, we would probably be able to create something indistinguishable from humans that could automate all possible jobs. If you think politicians or CEOs would be the most difficult, guess again. These people are essentially players chosen by a group of people to play a large and complicated game on their behalf (see the anime “No Game No Life” which takes a quite literal view of this subject). Computers are already significantly better than we are at chess, and the corporate/political game is not so different. It just has different objectives and pieces at play. Even if these positions may not be the easiest to automate, they are not the hardest. Automating coders also wouldn’t be the hardest — low-level coding work is pretty routine and we can expect low-level coders to be replaced in the next few decades.
In my opinion, the hardest jobs to automate are teachers, psychologists, artists and entrepreneurs. These are jobs that would require significant advancements in artificial intelligence and natural language processing to be replaced by robots, whereas most other jobs could be replaced by computer programs today if we really wanted to. Teaching people effectively requires an understanding of human thought, and teaching is also important if we’re going to get people to do anything machines can’t learn themselves. Psychology is important for maintaining the mental welfare of our society and, by definition, requires a deep understanding of human thought — it’s also something to which humans respond much better coming from a real person, although perhaps this may change in the future. Art can only be produced by robots if they can recognize the difference between good art and bad art, and most humans disagree on what this even means. Entrepreneurs and innovators are sort of shaky. While a good AI could theoretically optimize a product ad infinitum, it is hard to say whether an artificial intelligence would look at a flip phone and come up with better flip phones forever or eventually come out with the iPhone. Some of the design process also requires artistic knowledge and knowing what people want/need, which is a complicated issue for AI to address.
So what happens then? We will probably have a society centered around the professions that have not yet been automated. It isn’t that hard to envision a society in which each child has a personalized tutor who teaches them everything they need to know to contribute to society in a meaningful way — a huge surplus of teachers would help make this possible. The current model we have for entrepreneurs is also likely to change, since it places undue risk on the entrepreneur and we want to encourage entrepreneurship. Having novel ideas may very well become a profession, rather than a personal venture. Although these jobs that are difficult to automate are not necessarily highly compensated or saturated today, it is hard to imagine significant societal difference in status between these jobs in a future in which a large surplus of resources will render any notions of scarcity (and value derived from it) somewhat moot. We, as a society, will be encouraged to level the playing field since each remaining job will be considered essential and worth encouraging – after all, the continued existence of these vocations in spite of automation can be taken as evidence of their importance. I feel that most people in our society enjoy at least one of these types of work: talking to people, teaching people, making art, or thinking obsessively about new ideas. Almost all people would probably be happy doing one of these jobs for a living, especially a good living.
But let’s address the most important thing here: what you can do to prepare for this change in society. The first thing we have to do is accept it. Robots will automate our jobs, and no amount of anti-robot legislation will succeed in eliminating robots, just like anti-drug legislation doesn’t actually reduce the amount of drugs. So rather than just considering salary when we choose what skills and professions we learn, we should look also at how likely the profession is to be automated within our lifetime. So learn some CS – programmers will be okay for a while because they’re doing most of the automation. Learn how to paint or write or make movies. Learn about how the human brain works and what you can do to help other people who are having trouble with theirs. But, most of all, learn to respect other people, because in the future it will be harder to hide behind a high-paying job and a family name. Our future tends toward equality, and you should be ready for it when it comes.