The robot apocalypse has become a well-worn trope that will elicit laughter more than concern. But there is a far more direct threat from artificial intelligence or AI: economic disruption. Technology can and has taken jobs away from humans. I first started taking this idea more seriously after watching CGP Grey’s short documentary Human Needs Not Apply. If you enjoyed that video, this book is the must-read follow-up. Economist and historian Carl Benedikt Frey provides a soundly argued and clearly written book on the history of technological revolutions and what lessons these hold for future job security.
What inspired Frey to write this book was a chance remark from a fellow economist after a conference where Frey had spoken on the potential impact of AI on jobs. Are these concerns not overblown? Is this not the Industrial Revolution all over again? People found new jobs then, they will now. They did, of course, but that glosses over a lot of subtleties.
There are two take-home messages in this book. First, whether technology is opposed depends on whether it will hit people where it hurts: their wallet. Second, as history shows, successful opposition requires the support of those in power. Without it, resistance is futile. This might seem self-evident, but what I found eye-opening is the distinction Frey makes between enabling and replacing technologies, as not all technology is the same. Enabling technology is usually a boon to workers, making them more productive or lightening physically demanding tasks. Replacing technology, however, usually meets with fierce opposition, as it makes people’s jobs redundant. As an example, where motorised vehicles were an enabling technology compared to horse-drawn wagons that improved the job of professional drivers, self-driving cars promise to be a replacing technology that could mean mass-unemployment.
These clearly formulated messages are embedded in a grand chronological narrative of technological development and its impact on human labour, providing a much-needed and mightily interesting perspective. Frey takes the reader from the agricultural revolution some 10,000 years ago through to the eve of the Industrial Revolution in Britain, which he examines in detail. This period of steam-driven mechanization and job replacement led into the Second Industrial Revolution around 1880 with the discovery of electricity. The book’s focus shifts to the experience in the United States, which emerged as a powerhouse of manufacturing. This time around there was less resistance as living conditions improved and incomes were levelled across the board. That brings us to our current era where automation has emerged as a new disruptive force and inequality has grown once again. Finally, Frey carefully considers the future, with the looming spectre of AI promising further upheaval.
“[…] successful opposition [to technology] requires the support of those in power. Without it, resistance is futile.”
Above brief sketch of the book’s outline fails to tell you just how fascinating The Technology Trap is. Frey shines a light on forgotten historical episodes and provides a book full of fascinating explanations and reasonable arguments. I will highlight three strands that struck me in particular.
For one, a question I had never considered, why did it take so long for the Industrial Revolution to happen? Consider that historically there was no shortage of technological ingenuity and technical skill. From the engineering feats of the Greeks and the Romans who were more interested in warfare than industry, through (yes, even) the Dark Ages that saw windmills and ultimately the printing press, to the many scientific discoveries and inventions of the Renaissance. Leonardo da Vinci was but one of many inventors who were centuries ahead of their time. We had mechanical clocks, telescopes, and barometers long before the Industrial Revolution. So why the wait? Frey convincingly argues that the answer can be given in one word.
This is the technological trap of the book’s title. As also shown in Sheilagh Ogilvie’s recent book The European Guilds, craft guilds shaped the economy for centuries. They looked out for their members and opposed progress, either legally or by force. Frey provides many examples of violent riots destroying early inventions. And guilds saw themselves supported by the ruling classes who feared social unrest, with royal edicts banning machines. It highlights one of the book’s central messages; that successful opposition needs the support of those in power.
“[…] why did it take so long for the Industrial Revolution to happen? Consider that historically there was no shortage of technological ingenuity and technical skill. “
This first changed in Britain. Yes, that mountain of coal on which it sat also helped, but was not the only factor. Increased competition between fledgling nation-states and the waning of craft guilds brought about a shift, with the ruling elites siding with inventors, allowing mechanization to take off. This was an era of labour-replacing technology which initially saw inequality sky-rocket as a shrewd few pocketed the profits. Understandably, there was resistance, but the machine-breaking Luddite revolts were violently suppressed. The narrative of them holding back progress took hold, which is why “Luddite” has negative connotations today.
The second, equally fascinating story is how we got out of this. The introduction of steam power in factories required higher skill levels to tend to these machines, allowing many jobless adult men to find work again. Meanwhile, child labour made way for childhood education. The arrival of electricity and the combustion engine continued the trend of mechanization, creating new, labour-intensive jobs. This was an era of enabling technology where wages grew and work floors became noticeably safer. Frey points out the breakthrough of what engineers call “unit drive”. Where early factories had a central power source (water, steam, and later electricity) that transmitted motion to machines through a tangle of shafts and belts, at some point it became possible to outfit each machine with an individual motor.
Frey’s narrative here is largely US-centric, focusing on for example the assembly lines and new factory layouts pioneered by the Ford Motor Company, and the numerous inventions that flooded domestic households. It was a time of unbounded optimism and progress that saw the rise of an affluent middle class. He reflects here on the tendency of economists of each era to try and formulate universal economic laws, such as Simon Kuznets did in the 1950s. Kuznets pushed the optimistic idea that technological progress brings initial economic inequality, after which capitalism self-corrects for this with time by creating new job opportunities. That seemed to hold until about the 1980s when automation changed the rules of the game again. The famous French economist Thomas Piketty considers this era of affluence a statistical anomaly and some even suggest that violence and catastrophes are the only mechanisms to have historically levelled the economic playing field.
“[…] the age of automation […] is not the Industrial Revolution all over again […] why do we not see machine riots now, with people smashing computers?”
The third and final theme I want to highlight is the age of automation, which Frey argues is not the Industrial Revolution all over again. This has not stopped some from drawing parallels between the two anyway, notably the economist Branko Milanovic who thinks every technological revolution is accompanied by a “Kuznets wave”.
Although there was a time when computers were human, they soon became, well, computers. This split the existing middle class, with a minority finding work in the cognitively more demanding new job sector of software engineering and programming, while the majority of blue-collar workers were displaced into less well-paid service jobs. I hasten to add that, for the sake of brevity, I am leaving out nuances here, as Frey does not claim that this is the only or decisive factor – it contributes to what is happening. So why do we not see machine riots now, with people smashing computers? One reason is political participation: the difference between then and now is that all adults have the right to vote. Make no mistake, there is discontent, and politicians play on these sentiments. Frey thinks this is an important contributor to the rise of today’s populism and polarisation. And we are not done yet, as AI is at our doorstep. Without giving hard predictions, Frey takes a very realistic look at current and future capabilities and limits of AI, and which jobs will likely disappear. He is not favouring a new technology trap by opposing progress, but as his book shows, progress always comes at a price in the short term, often spanning the decades of someone’s working life. With the benefit of hindsight, governments can and should prepare for this, and Frey concludes with some policy recommendations.
One of the reasons I recommend The Technology Trap is its clarity, thanks to its excellent structure. It reads as a logical whole, each part starting with a helpful overview and ending with a summarising conclusion. Even to someone like myself, with a background in biology, this book is accessible. On top of that, it is simply fascinating. Rather than a dry treatise or “just one damn (historical) fact after another”, Frey brings to live the different periods he discusses. The blurb reeled me in with its interesting premise, and the book held me captive until it was done telling its deeply researched story.
The threat that technology poses to job security is of significance to most of us. If you want to understand technology’s impact, judge predictions about the future of work, or feel better informed when discussing these topics, you need to read this book.
Disclosure: The publisher provided a review copy of this book. The opinion expressed here is my own, however.
You can support this blog using below affiliate links, as an Amazon Associate I earn from qualifying purchases:
Other recommended links and books mentioned in this review:
CGP Grey’s short documentary Human Needs Not Apply