Everyone is talking about AI taking jobs. I believe that AI is going to change jobs, not take them, and it will take a lot longer than most people expect.
I empathise with the sentiment of AI taking jobs: ChatGPT blew people away and people assume large parts of life will be automated. As someone who’s neck deep in AI, the tweet below summarises my current position on AI:
To back up my claim, I decided to go back in time and review the largest technological innovations in history.
The past is not always the best predictor of the future, but it’s better than nothing. Finally, this is not a piece about AI regulation. Claiming that jobs will change does not negate the need for regulation. That’s a topic for another day and time.
Okay, let’s dive in.
Technology innovation
Humans have been obsessed about making things more efficient for as long as we’ve been around.
Brittanica has an excellent timeline of technology advancements. The full list is available here. I used it to compile the table below:
The colour coding is very broad stroke so please forgive me:
- Agricultural age (blue): we moved from hunter gatherers to cultivating crops, domesticating animals and living in settlements. More people were able to eat, and our population grew.
- Industrial revolution (yellow): we built machines that dramatically improved manufacturing, transportation and communication. In a purist sense, this was the first “phase” of automation. You could take a train to commute large distances instead of having to ride a horse.
- Computing revolution (green): most of the inventions in this phase were digital (computer, mobile phone, internet). There are two exceptions of course: nuclear power and spaceflight.
- To be named (purple): it’s hard to put a tag on this phase. Whilst everyone is talking about artificial intelligence, CRISPR and innovation in biotechnology is moving equally fast — you just hear a lot less about it.
When I was writing this essay, I spent 2 hours looking at that table. It’s incredible. So many inventions packed into a table made on a Google Sheet. It’s a fraction of what we’ve really accomplished. There are so many inventions I’d love to add: antibiotics, radiation etc.
Anyhow, it sets the scene well to think about how technology innovation has changed the social and employment landscape. Here are 3 inventions that really changed how we live.
The impact of tech inventions on society
The steam engine
The steam engine is one of the greatest inventions in history.
The steam engine also changed how people worked. Prior to its invention, people relied on animals, wind and water to power equipment. Once the steam engine was invented, the amount and type of labour required for this type of work fell.
This did create upheaval for some time. People lost jobs they had held for ages. The Luddites, for example, opposed the use of machinery because it replaced skilled labour.
Over time, the demand for labour cropped in other areas such as factories. It led to the birth of cities because factories were located in urban areas. This was also catalysed by the steam engine because it reduced travel time dramatically: travel speed went from 6 mph to 21 mph around London. For the first time ever, people could imagine their home and place of work being in different places.
My main takeaway from this: automation takes away jobs and creates new ones. The longer the gap is between these two, and the harder it is to train for the new job, the more people will oppose it.
Automobiles
The first automobile was invented in 1886 by Carl Benz. It was powered by a gas engine.
The credit for the first road trip goes to Benz’ wife:
“without her husband’s knowledge, Benz’s wife Bertha and their two sons Eugen (15) and Richard (14) embarked on the first long-distance journey in automotive history on an August day in 1888.”
Automobiles were being mass produced by the 1920s. Automobiles reduced the demand for horse carriages and put many of these folks out of business. On the flip side, by the 1950s, the automobile industry was the biggest employer in the United States.
Automobiles enabled private transport much like the steam engine enabled private transport. If you owned an automobile, you could choose to travel on your schedule much like Bertha Benz did.
But automobiles did something else: they created massive demand for roads. People were suddenly being employed to build highways. The most interesting aspect about automobiles and highways is how long it’s taken humanity to build them. Developed economies have great roads, developing economies do not. This general trend, where developed economies adopt technology faster, is a common trend.
There are two takeaways here. First, technology innovation creates complimentary jobs (e.g. roads for cars). Second, the more expensive the innovation is (in real terms), the longer it takes for people to adopt it.
The mobile phone
When the developed world thinks about the mobile phone, they think about the computer in a pocket. The real value of the mobile phone came from giving billions of people access to the internet.
People in developing economies couldn’t afford a computer but they could afford a mobile phone. This change is so dramatic that I’m struggling to put it into words.
Look at the graph below, which is a stark contrast to the adoption of automobiles. Can you think of a technological innovation where a developing economy in India has seen adoption as quickly as other developed economies? India is roughly a decade behind developed economies like the United States.
The mobile phone and internet did things for developing economies is hard to comprehend. Take India’s unified payment interface: who would have imagined that India would process 8.9 billion transactions in a single month (April 2023) electronically? I was in India 15 years ago and I’d have called bullshit on you if you had claimed this was possible.
My main takeaway: technology innovation is not uniform. It has dramatically different outcomes for developing and developed economies. In addition, most people are not good at predicting what’s possible 10 years from now.
The takeaways
- Innovation compounds: look at the table above on the big technology innovations that have happened. The start and finish of each of those periods has gotten shorter every time. For example, we did not need computers for Spaceflight (I checked), but I can bet NASA can’t imagine living without them.
- Penetration takes time: if it’s truly new innovation, penetration can take time. Automobiles were invented in 1886 but it’s taken nearly a century for them to become mainstream across the world.
- Progress is not equal: developed economies will benefit first in nearly every case. But the value innovation can bring to developing economies can be much higher. It opens up opportunities that could have taken decades.
- We don’t know what we don’t know: how many people knew that the invention of the transistor in 1947 would lead to the mobile phone, the internet and eventually artificial intelligence? Not many. Most of us are not built to see a future 100 years from now, and we have no idea what technology could enable.
- Jobs change more than they disappear: when technology changes, jobs change; they don’t just disappear. It’s the gap and our ability to adapt to these new jobs that creates unrest for people.
We adapt
The need of the hour is our ability to retrain. And we have never had better tools at our disposal for training and education.
For millennia, we have adapted to change. We’re actually really good at it. Today, we have the ability for people to learn according to their individual ability.
As a race, we are from not having things to do. It will take years for our jobs to fully change. Let’s use that time to prepare, learn and adapt like we always have.