How did society change during the industrial revolution? Are there lessons we can learn for the AI revolution?
You can listen to this podcast on Spotify, Apple Podcasts, Substack or wherever else you get your podcasts. Or you can watch this conversation on YouTube.
If you’ve been paying attention to the AI debate, you’ve probably heard some version of…
“I would love to compress what happened in 200 years of the industrial revolution into a 20 year period.” Satya Nadella on a recent episode of the Dwarkesh podcast.
But what does that actually mean? Does it include the mass social unrest and stagnant real wages that accompanied the industrial revolution in Britain? Are we focusing too much on where society ended up, and missing the reality of living through technological disruption?
In this episode of Ideas in Development, we try to tackle these big questions by asking whether the industrial revolution is a sensible comparison for the AI era, and exploring what that would imply. To do so, we are joined by economic historian Bruno Caprettini, who has researched the impacts of new technologies during this period.
Setting the scene: England before the take-off
Living standards in many parts of the world today would have been inconceivable for anyone living in 1400, 1500 or 1600. And it all started in England between 1700 and 1800, which has been a magnet for economic historians.
At this time in England, during the 1700s, most people lived in the countryside and worked the land, typically on relatively large farms owned by someone else - this is an important distinction with smallholder farms in some developing countries today, which tend to produce food for their own consumption. This work was (and is) hard and seasonal.
At that time, the North was much poorer than the South, with Manchester home to less than 10,000 people, whereas London already had about half a million inhabitants.
The Napoleonic Wars, labour scarcity, and incentives to mechanise
Britain was involved in many wars in the run up to the industrial revolution during the 1700s. Then came the most significant, the Napoleonic Wars of the early 1800s: at this time, around 10% of the British prime-age labour force were either in the Army or in the Navy. This was a major labour-market shock. With so many men away, wages rose, and employers had stronger incentives to substitute labour with machinery.
This ties into one of the classic debates in economic history, why industrialisation began in Britain. There are two main stories:
- Relative prices, often associated with historian Bob Allen: Britain had high wages and cheap coal, making it profitable to adopt coal-intensive, labour-saving technologies.
- Human capital (or skills), often associated with economic historian Joel Mokyr: Britain’s edge was the quality of its workforce, particularly the mechanics and skilled trades that were able to build, adapt, and maintain new machines.
Bruno’s work shows that rather than competing, these two hypotheses are actually complementary. By looking at where labour scarcity was most intense (using recruitment patterns for the army and navy as a proxy) and where mechanical skills were more abundant (using apprenticeship patterns in mechanical trades), Bruno and coauthors’ find that both mattered. The fastest adoption happened where both high wages and local mechanical capability overlapped.
Luddites, unemployment, and the Captain Swing riots
What did it feel like living through this period? Bruno draws a distinction between the war years and the decade that followed. During the war, food prices rose steeply. For those rural workers tied to agriculture, high food prices were actually beneficial, because they supported farm incomes and labour demand. A saying at the time went something like:
To another wet winter and one more year of war
After the war, however, the situation changed quickly. Soldiers returned home to find a labour market transformed by mechanisation. Wages fell, unemployment rose, and living standards deteriorated. Bruno describes the 1810s and 1820s as years of intense social unrest: urban textile workers broke machines in the Luddite movement, and rural areas later saw their own backlash.
This culminated in the Swing Riots of the 1830s, a wave of rural unrest often described as the largest episode of social disorder in English history. Bruno stresses both how widespread the riots were and how frightening they seemed to elites (the French Revolution was within living memory for older elites).
During the riots, Captain Swing’s name was attached to threatening letters sent to farmers, warning that their machines would be destroyed. Historians have proposed many explanations for this unrest, including a resentment of Irish migrants, anger about the poor rate (a rudimentary welfare system), and broader political grievances. But Bruno and his co-authors find that the machines themselves, specifically the threshing machines that reduced winter employment, were a central catalyst.
Why were threshing machines such a flash point?
Threshing is the process of separating grain from harvested stalks. Done manually, it was labour-intensive winter work, often requiring large groups of men beating crops for days on end. But during wartime labour shortages, farmers started to adopt machines that could do the same work far more efficiently, and the machines could be operated by women and children.
Bruno describes how, even when labour became cheaper after the war, farmers felt unable to abandon these new machines. Competitive pressures created a race to the bottom: if your neighbours used machines and you didn’t, your profits fell and you risked being outcompeted.
The result was not always a clean ‘wage-adjustment’ story. In many areas, wages were already close to subsistence, leaving little room for them to fall further. The adjustment showed up instead in job loss and underemployment, particularly during winter months when threshing had previously provided a crucial income stream.
“They couldn’t reduce the wages further because there was just no more space to go, but they made a lot of people unemployed.”
And what does all of this have to do with AI?
Two features made threshing machines politically consequential: scale and speed. In their research, Bruno and co-authors attempt a back-of-the-envelope calculation of how many jobs the machines displaced. The estimate is stark.
“We think that the machines were taking at least one third of the jobs of the people that were employed before the Napoleonic War in the countryside.”
* Remind you of anyone’s prediction? (AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10-20% in the next one to five years, Amodei told Axios in an interview.)
In the industrial revolution, this happened rapidly. The machines arrived during the war, so by the time workers returned, there was no gradual transition.
Bruno sees parallels in AI: rapidly expanding capabilities and the potential to affect many roles quickly. But he also highlights some key differences.
For one, AI appears to target much more advanced skills than the mechanisation of the 18th and 19th centuries. Whilst industrialisation did replace some skilled artisanal trades, many are predicting that AI could (eventually) automate large swathes of highly educated professionals, including lawyers, accountants, or even PhD researchers. That scenario would create a distinct political economy problem.
Second, the majority of societies now are richer and have larger welfare states than 1830s England. This will, to some extent, cushion income shocks. But Bruno argues that income support alone won’t be enough, because jobs also offer an identity and a purpose.
One key finding from Bruno’s work is that disruption was less damaging in villages near booming industrial towns. Proximity to alternative employment made adjustment possible: rural workers could migrate to manufacturing jobs, and within a generation, manual threshing could disappear without leaving permanent devastation.
Where does that leave us?
If AI truly is an industrial-revolution-scale transformation, history suggests we should expect not just productivity gains, but intense distributional conflict, contested politics, and a high-stakes transition period. The path societies take will depend as much on institutions and labour market dynamics as on the technology itself.
So if you are a tech leader who plans on invoking the Industrial Revolution in your next podcast appearance, please be clearer. Are you talking about the eventual prosperity, and not the decades of unrest and hardship that preceded it? Can you achieve the long-run gains without the short-term pain?