The European: Almost eighty years ago, the great physicist Nikola Tesla mused that “in the twenty-first century the robot will take the place which slave labor occupied in ancient civilization. There is no reason at all why most of this should not come to pass in less than a century, freeing mankind to pursue its higher aspirations.” Do you share Tesla’s optimism?
Steiner: With any large trend that is changing our lives, it’s never quite as simple. Demonstrably, algorithms are changing our lives quite radically. Some of those changes are good, and some of them are bad, especially if you happen to be working in a job that could soon be replaced by algorithms. You might be able to re-train and get another job, but your position in the economy will basically be eradicated. That’s a tough spot to be in.
The European: We tend to think about unskilled labor as the most precarious form of labor – machines could easily do it. Yet one of the arguments you make is that algorithms threaten many mid-level, white collar jobs…
Steiner: Being an expert in a field, having worked for fifteen years in that field, usually means that you have accumulated enough expertise, seen enough cases, read enough studies, dealt with enough clients, that you develop your own pattern recognition system within your brain. In medicine, experienced doctors are valuable because they have seen and treated many patients and diseases. Experienced lawyers know very well what information they need to pursue litigation, where to find that information, and so forth. But algorithms are very well suited for pattern recognition, much more than humans. If you can feed algorithms with data about a patient’s symptoms or about a legal case, I can’t see how that would not take away many of our jobs.
The European: What kind of work cannot be automatized?
Steiner: The kneejerk response is to say: ‘Things that take creativity and innovation.’ That’s true to some extent, but algorithms can be creative. In many ways, creativity means incrementally building on past innovations, and there’s no reason by a machine could not do that.
The European: I can’t see literature prizes being given to machines…
Steiner: Algorithms aren’t at the point yet where they can write a novel. They can certainly write prose and any kind of data-driven stories – about news, about sports, about the stock market – and will probably replace human workers in those fields within the next ten years. Humans would only have to contribute personal analysis or colorful observations. In medicine, some work will always require a human hand, but many tasks can probably be automatized. What this means is that the people on the high end of the food chain will be more empowered. The people who should really be afraid are those who do their job just good enough. You will see that in every field.
The European: You’re talking about a momentous shift in the labor market: The vast majority of workers fall into the category of mid-level employees. Unemployment is already high in the wake of the crisis – and if you’re correct in your predictions, we are looking at a technological shift with tremendous social consequences.
Steiner: Economists go back and forth on this question. Some argue that unemployment will remain high because of these trends, but others point out that people had very similar worries during the industrial revolution. The steam engine revolutionized production, but you still needed as many workers as before. The biggest trend for me is about stratification: The people who can handle algorithms and market themselves will do very well. They will displace people with less expertise and experience.
The European: Does that worry you?
Steiner: I’m more curious than worried. If you look at the top 500 American companies, corporate profits are near an all-time high, yet unemployment levels are higher than average as well. Many companies have not hired back the workers they laid off but have instead managed to increase productivity through software and automation. This isn’t just a future trend, it’s already happening.
The European: We are now reaching a state in which human labor isn’t necessarily the driving force behind economic growth anymore. Do we have to rethink our approach to compensation – for example, by redistributing the proceeds of automation to account for a loss in wages?
Steiner: I would be reluctant to do anything right now, because it is very hard to stand up and say something that would discourage tech innovation. The West has lost a lot of ground internationally to other countries, but we certainly have not lost ground in technology. A lot of the companies that are now heading the stock indices are rather new: Google, Apple, etc.
The European: But the problem still stands: How do we re-think the labor market if the industrial economic model of the past vanishes?
Steiner: I would argue for a more progressive and simple tax code. I don’t care if someone makes money in the oil business or in the software business: If you make a billion dollars, you should be taxed accordingly. I don’t want to penalize people who work on automation, because that’s where growth lies.
The European: If automation turns out to be such an influential trend, it appears possible to run our economy on a twenty-hour workweek. Yet the basic narrative today is that “those who work shall eat,” and that idleness outside of a structured work environment is unproductive. Is that narrative sustainable?
Steiner: I am not sure whether the number of jobs will decrease or whether we will be looking to fill different jobs. In any city, the job market for programmers is very good because there are many more jobs than people with the required skill set. We will still need a physical world of work around us: that is generally what makes us happy and powers our lives. Software is leading us towards greater wealth polarization, but I don’t think it is leading us towards a world in which fifty percent of the population sit on the bench. One narrative that has to be abandoned is that some people aren’t as quantitatively inclined as others. It’s true to a certain degree, but it’s often just a stereotype, especially for women.
The European: Some of the work that seems hardly replaceable by machines – for example, in the care economy – isn’t very well paid or valued. Do you think those profoundly “human” jobs will rise in prestige as a result of automation?
Steiner: One reason robots won’t replace those workers is not because they couldn’t do the job but because human labor is cheap if it is not skilled labor. Robots are expensive, so why wouldn’t you pay someone 40,000 dollars to work in the care economy? That’s why the people whose jobs are most threatened aren’t at the low end of the spectrum but in the middle.
The European: But it’s very cost effective to produce products with machines instead of human labor – which is why some economists are already talking about a resurgence of Western manufacturing. If you’re running a factory of robots, it doesn’t matter too much whether it’s in India or Indiana.
Steiner: We can already see that trend: Manufacturing companies are moving back to the United States. Another factor is the cost of energy: It used to be very cheap to ship a container of goods from China to the United States, but the cost of energy has risen steeply over the last three decades. Why pay for this giant shipping infrastructure if you can make the same things here for only a little bit more money?
The European: Is the time of global supply chains over?
Steiner: The more value a manufacturing process adds to a product, the more able it is to be shipped. But there’s another issue: In the United States, we simply don’t have the know-how to manufacture something like an iPhone. We design it, but China manufactures it. There are probably ten Chinese companies that could build up a production line for a new smartphone in less than two weeks. There is no company in the United States that could do it.
The European: One difference to earlier cycles of innovation is speed: It took decades for the industrial revolution to unfold, but it took only a few years for algorithms and automation to become driving forces. And because of that speed, our political and social arrangements are lagging behind economic realities.
Steiner: Politics can hardly keep up with anything. The software laws in the United States are broken, for example. Partially it’s a question of political competence, but it’s also incredibly hard to legislate. And it is made even harder by the speed of technological and economic trends. That speed is also the reason why companies like Apple or Google will not be able to stand on top for fifty years like Rockefeller did. You cannot be the top innovator for long. Look what happened to Microsoft’s empire: It took only ten years for them to crumble. First came the web, and then came mobile, and it is pushing faster and faster speeds. The limiting factor will not be the speed at which companies and innovations can be created but the speed at which customers can absorb these innovations.
The European: One argument about the downsides of technology goes like this: The fault is not in our gadgets but in ourselves. We build algorithms and machines, and they merely reflect our desires and aspirations. Given the speed of innovation cycles, do you think we’re actually in control?
Steiner: Nobody is steering anything. You have thousands of companies trying to solve their own problems – to grow and to make money – through software. That can lead to problems: Wall Street became a complete Wild West and investor interests aren’t always held in the highest regard. Many people simply did not know what was going on, especially when trading started to be computer-based. You saw the capacity for greed becoming menacing. But the situation was probably worse twenty years ago.
The European: It was worse because the rise of algorithms eliminates the potential for human error and emotion? During the 2011 “flash crash,” computer-based trading caused a market collapse, but also self-corrected within a few minutes.
Steiner: One of the core competencies of algorithms is risk control, but it’s only as good as the programmer. It can only react to freak events that it has encountered before, and based on the rules it was given. The problem is that lives can be ruined and billions can be lost within seconds. That is scary. In the next ten years you will see a movement for cars driven by computers and I’m afraid that it will never happen because people are scared. That’s ridiculous! There will be accidents and risks, but on balance those cars will be much safer than cars driven by humans.
The European: In your book, you quote the data scientist Jeffrey Hammerbacher, who said: “The best minds of my generation are thinking about how to make people click on ads. That sucks.” How much of machine-driven growth and work would you characterize as substantively “meaningful”?
Steiner: The brains tend to flow where the money is. A lot of good minds are working on banner ads nowadays. It’s similar to the trend in the 80s and 90s: A lot of money was lying around on Wall Street, so you got a lot of smart people who wanted to work there. But I believe that the trends will broaden: You can already see it in medicine, and I’m hoping to see it in other countries as well, for example in India. They used to do a lot of consulting programming for others, but they are now starting to make their own apps. We’re seeing the democratization of digital skills around the globe.
The European: Many countries eschew Fordist processes of industrial development to jump straight into the data economy…
Steiner: It’s clearly happening. Why wouldn’t they? The big question is whether that trend will create more countries that are more economically stratified? The people who know how to wield and create software will rise to the top of the food chain and those who don’t have that knowledge will be increasingly reliant on welfare money. We have already seen that effect in the United States. The country is becoming more unequal, and software has something to do with that.
The European: Our educational system is still predicated on the idea of preparing people for manual or traditional office labor. Do we need to teach and learn different qualifications?
Steiner: The most important thing is that everyone gets an equal chance. Technology should be taught early, programming should be taught like we teach math today as a core competency. That doesn’t mean you have to be a programmer, but nobody expects you to become a mathematician because you studied math in third grade. Why are social studies more important than technology? I’m not against culture, but to deny some kids the chance to be exposed to technology early on is detrimental to their chances in a very competitive global economy. It’s a mistake.
The European: I think I’m still concerned about the disparate impact of innovation: Some groups will inevitably benefit, and some groups will inevitably suffer.
Steiner: You can’t fix everything with education, but it’s a start. Give everyone a chance to prosper in this economy. But you’re right: Innovation is a social issue. If the gap between the haves and the have-nots continues to grow, we need social policies to address that.
Did you like the conversation? Read one with Hiroshi Ishiguro: “Robots help us understand human nature”