It is impossible to have a complete discussion on the impact of AI and Machine Learning without first considering History. In the series introduction two points were raised about AI and Machine Learning. First that its current focus is advanced automation, and second that this automation will revolutionise our lives, our work and economy.
History is a very useful tool for assessing what the future holds. Little is truly new in the human experience and the past offers examples of how humans responded to similar problems to the ones we face today. Automation is no different, it is not a new phenomena and has a long and interesting history.
The roots of automation took hold in 18th century Britain and then expanded across Europe and America. During the 1700s a great number of inventions and discoveries began to increase human power and efficiency, and enable automation.
In 1712 Thomas Newcomen developed the first commercial steam engine; by the end of the century inventors had begun to experiment with internal combustion engines; In 1821 Michael Faraday unveiled the first electromagnetic rotary motor; Fast forward a century and in 1943 Tommy Flowers had built the first digital, programmable computer; In 1969 the first message was sent across ARPANET; and by 1993 thanks to the work of Tim Berners Lee we had HTML and the birth of the Internet as we know it today.
Across a period of roughly 300 years automation flourished, driven by invention. Of course the people referenced here are merely a selection of those involved in the great automation revolution. Also many of the inventions highlighted were not truly original discoveries, the steam engine for example may date back as far as the 1st century AD, and humans have played with computation tools as far back as the second century BC. However during the 1700s there was a confluence of invention that was put to commercial and industrial use in a way not previously experienced.
There are though two people not already mentioned but of great importance to the story of automation. They weren’t inventors as such, or maybe not true engineers or scientists like the others mentioned, they were though two of the great entrepreneurs and industrialists who helped drive automation forward. They are Richard Arkwright the father of the factory system and Henry Ford the father of the assembly line.
Richard Arkwright was born in 1732 in Lancashire England and made his name and fortune in the cotton industry. In fact it would be fair to say he created the cotton industry. In 1769 Arkwright with the help of John Kay developed the spinning frame, which later became the water frame. The spinning frame automated the process of turning raw cotton into thread, which greatly reduced its production cost. Arkwright also worked on improving carding machines that turned cotton buds to cotton fibres.
In 1771 he set up his first water powered factory in Cromford Derbyshire. As a result Arkwright was probably the first person to mechanise and automate the process of turning a raw material into a product with little human skill required. Arkwright grew an empire, by the time he died in 1792 up to 30,000 people were employed in factories using his patents and he left a fortune of £500,000.
Henry Ford was born in Michigan, USA in 1863, in his early career he was an engineer and worked for the Edison Illuminating company. It was after this role that he began his journey towards founding the Ford Motor Company. It took four years, two business failures and a lot of help from investors before Henry Ford founded the Ford Motoring Company in 1903.
Ford’s great innovation came in 1908 when he released the Model T which was to be the first mass market car and triggered the automobile revolution of the 20th century. When sales began the Model T sold for $825 which was cheap for the time. Initially production was slow by 1910 they had only produced 12,000 vehicles. However after moving to a new factory and innovating with the assembly line and use of machines Ford managed to reduce production time for the Model T from 12.5 hours to 1.5 hours. By 1925 production had hit 10,000 vehicles per day, and by 1927 when Model T production ceased over 15 million vehicles had been produced and the cost had fallen to $260. It wouldn’t be until the 1970s that another car surpassed this production record.
It should be noted that Henry Ford did not invent the automobile nor even the assembly line. He and his engineers though did refine and enhance both to such a level that they triggered a manufacturing and automation revolution that we enjoy the benefits of to this day.
The story of the automation revolution is one of great interest, it is also far more complex, intriguing, and involves far more individuals, groups and companies than I have been able cover in this short post. We are though left with a very important question to answer, how has this automation revolution effected us?
Since the 1700s and the dawn of automation economic wealth has grown exponentially. Between 1990 and 2010 extreme poverty was reduced by 1 billion people and in the last 30 years China’s economy has grown from $300 billion to $11 trillion.
We should have all seen the chart below it tells the odd story of how humanity for a very long time made little to no economic progress and then in the 18th century things began to change dramatically.
There can be little doubt that automation has been a great benefit to humanity and has improved the lives of billions. This should make us hopeful for the future, automation could make this picture even rosier.
However this path to economic prosperity hasn’t been a smooth one as the chart suggests. There are many questions regarding the impact of automation on employment and concerns it may put many people out of work. These concerns are so significant that Elon Musk has suggested we may need to introduce a universal basic income.
One of the key concerns with automation is the impact it may have on employment, and therefore livelihoods. Marx raised these concerns in the 1860s in his seminal work Capital.
“The instrument of labour, when it takes the form of a machine, immediately becomes a competitor of the workman himself.”
Historically there is no clear relationship between automation and unemployment though. Currently in the UK unemployment stands at 4.4%, April to June 2017, and it is estimated that in the 18th century unemployment stood around 10%. Over the last few hundred years unemployment has dipped below 4% and grown significantly higher than 10%. For the UK we have data that shows the unemployment rate since 1881, and what can be seen in the chart below is that unemployment has spiked at points.
Unemployment though has not lasted and become a permanent issue. The causes of these unemployment spikes are also complex and difficult to pin down. It is unlikely, or at least there is not a great deal of evidence to suggest, that automation is directly linked to these unemployment spikes.
It is also unlikely that automation has had no impact on unemployment or employment. What history suggests is that automation is a disruptive process that changes employment trends, it changes what jobs people do. It is part of the process that Joseph Schumpeter described as the “perenial gale of creative destruction.”
Richard Arkwright’s factory drove a great deal of employment, but this was for unskilled labour. In the decades after Arkwright we saw the rise of movements such as the Luddites. These movements were not a reaction by the poorest members of society against the evils of the new factory system though. The Luddites were skilled workers who feared that the rise of automated factories manned by low skilled labour would threaten their livelihoods and place in the productive process. They were right of course and their attempts to halt this change ultimately failed.
As the Ford empire began to grow during the 1910s based on significant improvements in automation we saw different results. In 1914 Henry Ford introduced the $5 daily wage, this doubled the average wage for workers, and by 1926 Ford also introduced the five day working week. While Ford was no saint when it came to dealing with his workers these were both significant improvements to worker conditions. They were based on a belief that to attract the best, most productive and skilled workers you had to treat them well and this in turn would boost both production and wealth. It was a paternalistic outlook but it was correct.
Another change was the growth of completely new industries, and tourism is a great example of this. In the 18th century tourism was barely a thing, reserved for the wealthy and religious pilgrimages. In 1841 Thomas Cook led his first excursion taking 500 members of the Temperance movement on a return trip from Leicester to Loughborough. By the 1850s he was organising excursions for tens of thousands of people and even began organising foreign trips. Today the Thomas Cook Group has revenues of over £7bn, and the tourism industry as a whole is worth over $1tr annually.
Despite the brevity of this post and its cursory glance at the history of automation it does highlight some important trends. Automation drives growth in wealth, it doesn’t directly lead to unemployment, but is a disruptive force. It leads to changes in employment trends that will have a negative impact on some workers and industries, but at the same time creates opportunities for new industries to grow.
It is likely that the future of automation, driven by machine learning, will be very similar. It will lead to greater wealth, but it will lead to unemployment in certain areas and some skilled workers will lose their jobs. It is though unlikely to drive mass, persistent unemployment as new industries are likely to develop and provide new opportunities for people to thrive.
History suggests we should be quietly optimistic about the automation revolution that is to come.