Spark the innovation within you.

When was the last time you did something new?


AI vs. Industrial Revolution: 10 Comparisons


You’ve probably read comments crowning artificial intelligence as a new Industrial Revolution. But is that truly relevant? Beyond their technological underpinnings, we wanted to carefully examine the striking similarities, as well as major differences, so as to better grasp the uniqueness of the ongoing AI boom.

1// Origins


Expanding available workforce => UNCLEAR

Thanks to a rise in food production and food productivity, the European population grew rapidly at the end of the 18th century (x2 in the UK between 1750 and 1800), and a smaller share of the population was needed to work in agriculture, thus leaving numerous laborers available for other tasks. Hence they ended up in factories created during the Industrial Revolution.

The situation is unclear today: in the last 25 years, there hasn’t been a significant uptick in the world unemployment rate. On the contrary, it has slightly decreased, from 6.1% in 1991 to 5.7% in 2016. Still, 2 billion inhabitants have been added to the world headcount during this period.

Are absolute or relative figures more relevant? Could the sheer size of additional labor force fuel the AI boom by participating in it (e.g. training algorithms), or find jobs indirectly through AI, within industries renewed by it? Our only answer for now is to wait and see.

Changes in customer demand => SIMILAR

Changing customer tastes were a driving force behind the Industrial Revolution that started in England. In particular, the demand for new, colorful clothes drove the incredible growth of the (cotton) textile industry.

Measuring the final demand for artificial intelligence is tricky, but it is not far-fetched to think that the digital revolution and its multiple experiences lead to a generalized demand for real-time, heavily-personalized, scalable services for which machine learning algorithms are inevitable ingredients.

Strong entrepreneurial spirit => SIMILAR

Several factors fortifying the “entrepreneurial spirit” in England, such as the property rights system, the development of financial markets or new legal forms of companies, have been credited as being some of the causes of the Industrial Revolution.

Entrepreneurship in its startup form (i.e. tech-oriented, with a “go big or go home” mindset) is flourishing: MBAs increasingly preferring to create their own companies rather than joining blue-chip ones, venture capital funds abounding, the emergence of unicorns (that is, beyond the hype, the possibility of funding private companies with tremendous funding rounds, with almighty founders at the helm), the fast-growing volume of code publicly available thanks to open-source projects and collaborative tools including GitHub… all are factors conducive to a healthy ecosystem to nurture AI-enabled companies.


2// Impact


At first, tough times for the workers => DIFFERENT

At the beginning of the Industrial Revolution, labor conditions deteriorated for factory workers: lower wages, jungle law-style layoffs, health risks, greater number of hours worked, erasure of traditional communities and their solidarity schemes…

I’m sure this opinion will be highly debated, but I don’t think we’re heading toward a new wild adjustment through the AI boom. Of course we can expect displaced low-skilled workers to only find low-pay service jobs; difficult adjustment to new workflows driven by AI algorithms etc. But to me, by no means is this different from the disruptions that have been brought about by the computer age in the last 50 years. Above all, the starting point is way different from the background of the Industrial Revolution, since there are numerous social and economic regulations in place.

The real difference - again, a continuation of the computer age - could be an increasing conflation of work and private lives (i.e. managing a personal smart assistant from work, or the rise of what some call digital labor - all the small tasks you do for free, such as tagging pictures on Facebook, which can be incredibly useful to train AI algorithms).


Increased job specialization => DIFFERENT

As is well-known, the Industrial Revolution resulted in a division of labor into specialized jobs, a prime example being the automobile production line.

Whatever one’s opinion about AI’s net effect on society (i.e. cornucopia or apocalypse), it should lead to decreasing specialization, because:

  • either AI will eat narrowly-specialized jobs,
  • and/or it will enable employees to perform more easily a wider range of tasks thanks to various AI tools.


Concentration of the workforce in factories=> UNCLEAR

The advent of the Industrial Revolution was the advent of a new type of organization, initially for practical reasons: coal & steam were concentrated power sources, hence they required a concentrated workforce to leverage them. Why has the workforce concentration subsisted, in factories and offices alike? Economies of scale (production but also communication of information) are probably the reason.

In the digital age, and especially in the AI boom, (data) network effects dominate. Distributed labor organizations (e.g. remote work schemes in many startups) are thus more plausible than during the industrial era, leveraging digital tools to maintain a coherent culture and efficient internal communication.

That being said, if a concentration of a given company’s workers is less necessary today, it doesn’t mean that factories disappear: in the era of AI, these are called data centers, with gigantic computing power to train algorithms and disseminate them on a global scale.


Concentration of the population in cities (aka urbanization) => UNCLEAR

During the Industrial Revolution, concentration didn’t only occur at a factory level; it also happened on a wider scale, that of cities: the phenomenon of urbanization, and the expansion of cities feeding on migration flows. Immigrants came from the countryside (e.g. 6 million African-Americans moved from the South of the US to the urban areas of the North between 1916 and 1970) and from abroad.

We’re at a crossroads now; two antagonistic trends are at stake. The first is the never-ending growth of metropolises: rise of the services sector, globalization, startup ecosystems… every transformation of the economy has reinforced the importance of large, dense urban areas aggregating cultural, educational, economic, financial and sometimes political functions. And the emergence of AI hubs in Silicon Valley or in Montreal could fit well in that trend. In addition, the possible disappearance of some categories of jobs could push displaced workers to move where the most employment opportunities lie.

The second trend is more uniquely connected to the AI boom: if it leads to a dramatic fall in the cost of certain products and services, in the context of the digital world’s connections proving more important than physical constraints, we could finally witness firms spreading their workforce more evenly over the globe. Such as highly-automated factories being established in rural areas.


Dualization of the world => DIFFERENT

A clear division among nations arose from the Industrial Revolution: a few industrialized countries, and the rest supplying them (with coercition and colonization if need be) raw materials and cheap labor.

Here again, the impact of AI is two-pronged. On the one hand, the production of AI is more likely to be relatively concentrated in a few hubs, but there are already several centers, emerging in different countries at the same time (“US vs. China” in headlines), whereas the Industrial Revolution arose first in England before gradually extending to Continental Europe and to America, then to Japan etc. The contagion rate is faster due to digital communications, open-source projects and cloud resources. So the AI leaderboard will depict a clear hierarchy, but much less inegalitarian than what Industrial Revolution initially brought about.

And on the other hand, due to the same factors just mentioned, the exploitation of AI will certainly be widespread, and so its transformative impact on many industries.


Difficulty to measure the impact => SIMILAR

Here is how the historian Peter Stearns describes the unfolding of the early Industrial Revolution: “In its early stages it may have little measurable impact on overall production rates, which are still determined by more traditional methods of work.” This statement depicts incredibly well my feeling about the current situation of AI.

If it lives up to its expectations and ends up disrupting a sizable chunk of the economy, we’ll take a look back a few years or decades from now, and realize that the indicators we used to assess the evolution of our world are no long operative. Indeed, the absence of significant productivity gains measured in the wake of the computerization of society may be due to a measurement difficulty. For instance, online free tools now are everywhere, and even if advertising or other revenue sources can be leveraged to measure productivity at a firm level, there are still additional gains in the quality of user experience that statistics are bound to miss.


A revolution feeding itself => SIMILAR

The Industrial Revolution initiated virtuous feedback loops. The possibilities created by the steam engine generated a revolution in transportation (steamboats and railways) which in turn increased the demand in raw materials and metallurgy; the revolution in transportation transformed the retail sector (more goods available, ability to create large national chains) which in turn increased the demand in transportation (for logistics) etc.

The AI boom is on the verge of following the same path, starting from a few sectors such as transportation or finance. These disrupted industries will ask for more and more AI resources, which could then serve to transform other industries.

2 characteristics of AI make it even more prone to feedback loops:

  1. An increasing use of AI-enhanced services will expand the volume of available data, thus resulting in an improvement of algorithms whose quality will further increase user numbers and user engagement.
  2. AI researchers are increasingly using AI algorithms to help them tune AI models to improve their efficiency - so this could become a quick and direct feedback loop.



So the net result of our investigation is… incredibly mixed. The Industrial Revolution and the artificial intelligence boom have quite similar societal origins, share the same “shapes” (cross-industrial impact, feedback loops, measurement difficulties) but in 2 centuries the environment (above all international trade and job market) in which the Industrial Revolution took place has obviously changed.

Indeed, the AI / Industrial Revolution comparison gets the most interesting when we contemplate how the AI boom could reverse some consequences of the Industrial Revolution - in a nutshell, making possible a “flatter” world, where the fragmentation between countries, regions and jobs is lessened.