The Luddites are Back

In the evening of November 12, 1811, a group of workers collected in the moors of Nottingham, England.  After 2 successful attacks on local industries, they wanted more. Tonight they were drilling and practicing to attack 3 more mills.  The new stocking frames and cropping frames, they argued, were taking jobs away.  If they were not destroyed, their jobs would be gone forever.

For next 3 years, the “Luddites” ravaged English industrial towns, leaving destruction in their wake. Named after a young apprentice, Ned Ludd, this movement is largely associated with people that oppose technological progress on the grounds that technology takes jobs away. They argue that these technologies must be stopped or else their will be massive unemployment.

Over the last two centuries, various “Luddites” have re-appeared in various forms, arguing and sometimes fighting technological progress. Today, a new batch of Luddites have emerged, this time to fight against artificial intelligence and advanced robotics. This time, ironically, some very prominent technologists are leading the Luddite charge, most notably Bill Gates. That’s right, the same Bill Gates that founded and developed the largest software company in history (up to that point), is now worried that AI and robotics will take your jobs away.

In a recent interview at Quartz, Bill Gates said “at a time of when people are saying that the arrival of that robot is a net lose because of displacement, you ought to be willing to raise the tax level and even slow down the speed of that adoption”.  While Gates undoubtedly takes a less aggressive form of Ludditism, he still succumbs to the same premise as other Luddites. The basic premise is that automation and technological advancement leads to unemployment and that something needs to be done to stop or slow those changes.

That’s not to say that massive changes in our economy may be eminent due to improvements in AI and robotics. For example, self-driving cars could displace millions of jobs in the transportation industry. There are currently over 3 million truck drivers in the U.S. Yet, within the next 10-15 years, there will likely be fully autonomous vehicles, making truck driving as an occupation obsolete. In other industries, advanced robotics could replaces workers by the millions as well.

Bill Gates is not the only one. Other technologists such as Elon Musk agree.  Physicist Stephen Hawking is also worried. And Noble Prize winning economists such as Paul Krugman also show sympathy for Luddites. But for every luminary that is worried, there are equal number that are not. Yet these others rarely make headline news because “the world is not coming to an end” doesn’t sell.

But Bill Gates is wrong to recommend a tax on AI and robotics.  Why? Three reasons – one historical, one economic, and one moral.

Historic case against Luddites

Automation has long been the focus of Luddite attacks.  New technologies automate the jobs previously done manually or radically reduce the amount of manual labor needed. Consider the textile industry.  The advent of spinning mule, the Cartwright loom, and the steam engine created the key technological advancements that automated much of the textile industry. While there were only 2400 looms in England in 1804, within 50 years, that number had blossomed to over 100 fold, many with more advanced innovations that could do more work with less manual assistance. With this massive increase in automation, it would be natural to expect a massive increase in unemployment across England. Yet, this was not observed. In fact, as Andrew Bernstein documented, that not only did unemployment fail to rise, but the England witnessed a massive growth in total population – employing everyone. Technology automation did not lead to widespread unemployment.

In the late 1800s and early 1900s, another explosion in business innovation, called the second industrial revolution. After massive drops in the price of steal, transportation, chemicals, and communication, businesses grew exponentially in size.  Automation of factories started in earnest, eliminating expensive employees. Every new automation should have meant the lose of substantial number of employees.  Yet, that was not observed.  Employment today is still hovering at around 5% in the U.S.  In fact, the wealth throughout he world is increasing.  Recent data from the World Bank says that extreme poverty has dropped below 10% world wide. But according to the Luddite premise, that should not happen.

Bill Gates is not a stupid man.  He is extremely well read and likely knows much of this history.  So why does he not find this history lesson sufficient to change his opinion.  It’s because he (and those like him) thinks this time it’s different. This time, they say, the changes will happen so quickly that the newly unemployed will not be able to find alternative employment because of the mass of unemployed will overwhelm the marketplace.

Economic case against Luddites

Unfortunately, the claim “that this time its different” still falls for the Luddite fallacy, which says that innovations lead to increased unemployment in general. It is true that rapid technological change can lead to short term job displacement. However, what’s true a localized, micro level does not necessarily lead to macro unemployment.  Why?  Because technology is only introduced if it reduces costs.  As the cost of goods decrease, so do the prices at which they sell (which occurs when competition is allowed to flourish).  This means people will have more expendable income for additional goods or services – creating demands in new areas and new jobs. For example, current projections for jobs in elderly care show huge gaps between the current employment and future needs.

Increased productivity does not relate to unemployment. As Economist Alex Tabarrock states: “If the Luddite fallacy were true we would all be out of work because productivity has been increasing for two centuries.” Tabarrock further points out research that shows productivity gains usually accompany boom markets, suggesting that technological automation is a net jobs creator. If our government creates policy around slowing or stopping this technological innovation, we will likely hurt the jobs creation mechanisms in our economy.

Its not as if self-driving vehicles are going to appear over night and all 3 million truckers are going to be put out of work at the same time.  It will be a 15-25 year process, if not longer. Capital must be raised. Infrastructure must be built. People must be trained in the maintenance and care of these vehicles. Plus, younger adults will avoid the industry knowing that there’s no future in it, leading to a natural matriculation of drivers changing careers or retiring. Many of the current drivers will learn about the technology improvements long before it affects them and will adapt.  There will always be a few that refuse to change. They are like the last few people that live in a former industrial town, even though the factory shut down years ago.  Those last few inhabitants refused to recognize or made excuses for the change in their conditions.  Now they are “stuck” there. I put stuck in quotes because inevitably, there are options, some of which might be painful and disruptive short term, but best for the individual long term. Should we create policy based on a few irrational souls?

Moral case against Luddites

It’s this last question where most of the disagreement stems. I am a firm believer in the power of people to think and solve problems for themselves.  I’ve worked with lots of students over the years, including some that struggle with simple decisions.  And yet, when push came to shove, a good one-on-one conversation with the student always got a point across. People are not inherently irrational.  Certainly some do make irrational decisions, especially looking at it from the outside.  But generally, they do what’s best for themselves – sooner or later.

Yet, Bill Gates and his ilk are suggesting that we create political policy as if people are inherently irrational. They see the role of government as paternal – acting as parents, whether those actions are wanted or not. It’s this idea that we should use government to force others to do what we (the elite) think they should do. That’s not to say that government shouldn’t be there to protect the innocent and preserve our freedoms.  It should.  But there’s a fundamental difference between protecting our rights and forcing behavior that may “help” someone. Its immoral to violate one person’s rights just to “help” someone else.

But here’s the really baffling part.  Bill Gates argues that we should tax AI.  Artificial Intelligence is sophisticated software that acts intelligent. There are many things we use today that could be classified as AI, such as Apple’s Siri or Amazon’s Alexa. And software is just an opinion – an opinion of what behaviors should be enabled or constrained, an opinion of what data to collect or ignore, and an opinion of how things should be processed. In essence, Bill Gates is arguing that we should tax an opinion – tax ideas. Is this really where Gates wants to go with his recommendation? Taxing ideas?  I can’t imagine a more profoundly immoral idea.  It punishes people for their ideas, the very thing we need to survive.

And then there’s the practical side of the policy. What are the constraints on what is a robot or what is AI?  If we tax such an entity, it must be clear what qualifies and what doesn’t. Does it have to meet certain requirements?  Who is to decide these requirements?  What if an entity comes close but doesn’t quite meet those qualities?  This approach to governance opens the portals to serious lobbying and potential abuse.  First companies will lobby for their software not to be included in the tax, but once it is, they will lobby to create barriers to entry so other software companies can’t compete. New technologies with slim margins may never get developed, in spite of the fact it might be extremely useful to society.  Think of robots that help the elderly. With our aging population, this is certainly a concern.  Yet, this tax might prevent the robots from even being introduced if the tax prices the robots out of reach, killing the industry before it starts.

I hate that Bill Gates, a one-time hero of mine, has sunk to this mentality.  As a leader of the software revolution, he championed the use of software to increase our productivity.  But now he proposes the opposite, to slow software productivity gains.  It profoundly saddens me that I have to write this post.

Software – the driver of information revolution – is radically improving our lives.  I want that revolution to continue. I want to see a future where AI and robotics free us from the drudgery of manual labor and boring jobs.  The sooner it can get here the better. But to do that, we need to put the Luddite fallacy to rest and to kill all of its offspring in whatever form it takes.