In November 2015, the Chief Economist of the Bank of England gave a speech where he said that, within 20 years, 80 million jobs in the US and up to half the jobs in the UK would be lost to robots and algorithmic software (collectively referred to as 'robots').
A 2013 Oxford University study backs him up. It predicts that, of the 700 jobs types it evaluated, half of them would eventually be done by robots.
We see robots in factories on TV, working impossibly fast. But the future affects not only blue collar jobs, but numerous white collar ones, like lawyers and accountants.
The idea that this is possible is a fringe one for now, but this is becoming reality very rapidly. The CEO of one of the UK's largest construction companies just gave a speech where she said that skyscrapers in London would soon be built by robots rather than people. Alison Carnwath, Chairman of Land Securities, a FTSE 100 company, said ""Five years ago I'd have smiled wryly if somebody had said to me that robots would be able to put up buildings in the City of London â I tell you we're not that far off, and that has huge implications."
The rate of progress in processing power is astonishing. Computers have now beaten the world champions at chess (1996), Jeopardy (2004) and the highly complex game, Go (2016). Experts at the time all predicted that it would take a decade or longer for computers to beat humans at these games, if at all.
THE THREAT FROM ROBOTS: Since the early days of industrialisation, skilled workers have feared machines as a threat to their jobs, either by replacing them or allowing semi-skilled, lower wage workers using technology to replace them. But the reality has been different. As technology has advanced it has helped grow overall employment and aided productivity. And technology has helped raise the wages of unskilled workers. But this trend is now changing.
This time it is different and the impact on society will be dramatic. There is a paradigm shift happening where machines don't aid in improving productivity, they are becoming the employee, blue collar and white collar.
A few actual examples:
In the 1980s, the textile industry in the US was decimated, as three-quarters of textile jobs went offshore. Now, a textile factory in Parkdale Mills, South Carolina has been almost completely automated, and it produces the same output as in 1980, but using only 140 people rather than 2,000. Today, this new automation technology is so efficient that it is competitive with even the lowest-wage offshore workers.
Wall Street used to employ 150,000 financial workers in 2000. That number is down to 100,000 as automated trading programs using high frequency trading strategies now account for the bulk of the volume on the stock exchanges. These programs use self learning algorithms to adapt their strategies and have moved beyond the control and even the comprehension of their programmers.
Top media outlets such as Forbes magazine now use unacknowledged analytical and writing software engines to create news reports using multiple data sources. The CEO of Narrative Sciences, which markets the Quill software, predicts that within fifteen years, 90% of all news articles will be written algorithmically. What happens to journalists, corporate researchers, etc.?
Self driving cars are on the verge of becoming an everyday reality on the roads. As the sharing economy grows and home and car ownership becomes less important to the millennial generation, the demand for cars will decline in the West. Automation and fewer cars means jobs losses for taxi and truck drivers, car plants, parking lots, gas stations, and the service sector dedicated to automobile repair and maintenance.
Amazon now uses over 30,000 Kiva robots (which can pick up weights of up to 700 pounds) for packing and shipping. And that is in just the first 13 of its 50 warehouses. Amazon intends to scale up the use of robots dramatically.
Study after study shows that algorithmic approaches now routinely outperform human experts in many fields. And robots continue to improve exponentially, humans do not.
Robots don't get tired, they don't have attitude, they don't slack off and they don't want higher wages and benefits. They don't strike or call in sick. They will work continuously and holidays, weekends and vacations are irrelevant. They don't play politics. Robotic systems present a combination of speed, precision and brute strength, which translate into much higher productivity.
Businesses prefer robots over people. As the CFO of Nike put it, the long-term solution to rising wages in developing countries is going to be "engineering the labour out of the product." He's talking about ultimately no workers being used at all to make Nike products.
BIG DATA AND AI: Big data, combined with machine learning and cloud computing, is promising to provide analysis and insights that were previously impossible and which will have a revolutionary impact on business, medicine, politics and the sciences.
A big potential impact will be on knowledge based occupations. Think about what many office workers do. They look at data and, after analysing what they find and drawing on their training and experience, they reach conclusions and make decisions or recommendations. They may produce reports that support their analysis. Machines can do that now.
Bill Gates, Steven Hawking and Elon Musk have expressed grave concerns about the growth of artificial intelligence (AI) systems as a potential existential threat to mankind. Musk has invested in Google's Deep Mind AI unit just to keep an eye on how it develops! As Hawking warns "humans, limited by slow biological evolution, couldn't compete and would be superseded by AI." He believes that "success in creating AI would be the biggest event in human history."
CURRENT TRENDS: The Bureau of Labor Statistics reports that in 1998, workers in the US business sector worked a total of 194 billion hours. In 2013, they produced goods and services worth $ 3.5 trillion, about 42% higher after adjusting for inflation. Total hours worked: 194 billion.
Higher output with the same hours worked means productivity went up, thanks to advances in automation and technology. But the wealth created is not being shared evenly and this is causing rising income inequality. In addition, the population grew by 40 million in this period, so it is clear that there is dramatic unemployment and underemployment.
The headline 4.9% unemployment rate in the US is deceptive as it excludes people who are willing to work but can't find a job after a year ('discouraged workers') and it counts part-time workers as employed even though they may be looking for full-time jobs. The labour participation rate peaked in 2000 and is now at the level it was in 1977.
Real wages have been stagnant over the last fourty years. Virtually all this wealth creation has gone to the top 1%, which is why there has been such a revolt against the establishment in the current US election cycle. Brexit expressed the same rebellion against the system.
But what happens when tens of millions of additional workers and more lose their jobs in the West? And robots will also take away hundreds of millions of jobs in developing countries. What will be the impact on society as vast numbers of people can't find work and slide into poverty in the developed countries? For the jobs that still remain, the amount of surplus workers will ensure that wages will decline.
Citigroup wrote a series of confidential memos to their wealthiest clients. They argued that the US is becoming a 'plutonomy' â a top-heavy economic system where growth is being driven by a tiny, prosperous elite who consume an ever larger proportion of what the economy produces. They advised investors to avoid buying shares of companies that catered to a rapidly dissolving middle class.
The irony of all this is that in 1949, Norbert Weiner, an MIT mathematician who established the field of cybernetics, warned about the threat of "an industrial revolution of unmitigated cruelty" where eventually machines would "reduce the economic value of a routine factory worker to a point where he is not worth hiring at any price."
Automation within the military can be seen in the drone program. The Shah of Iran faced a popular rebellion in the streets in 1979 and ordered his troops to shoot to kill. But the soldiers refused to fire upon their fellow countrymen. Robot soldiers are already being tested. If they are ever deployed by dictators and other authoritarian leaders, robots will have no hesitation in shooting whomever they are ordered to kill.
As this robot revolution grows, the owners of the companies employing these robots will see their labour costs drop dramatically and their output increase substantially, leading to a massive rise in corporate profits and wealth. The rich will get much richer while the average standard of living falls for the rest.
And as technology creates new industries, it will also create new robots to service those industries. So innovation is unlikely to lead to new mass employment opportunities.
Robots will be of help to people where the work is dangerous, such as mining, explosives handling, deep sea and space exploration. So they will also have a legitimate public interest role.
PUSHBACK: A recent OECD report, conducted by research firm Forrester, challenges the potential for serious job losses. They peg the potential job loss at 9%, as they assert that many jobs require cooperation, which robots cannot do.
Yet even this optimistic report predicts that 12% of the 89 million white collar jobs will be lost by 2024, mostly office and administrative positions, and business salespersons. That is almost 11 million jobs!! It then expects that, by 2021 (just five years from now), cognitive systems to have reached a level of sophistication that allows meaningfulÂ participation in decision making, affecting management and finance positions.
Whether it is 9% or 50% of jobs lost to robots, the net job losses will be in the tens of millions in the US (and much more worldwide), with a huge impact on society in terms of unemployment numbers, reduced tax base and demands for increased government benefit payouts. This will translate into higher crime rates and the potential for social unrest.
SOLUTIONS: It is difficult to envisage meaningful solutions. Governments usually advocate education and training as remedies for those who lose their jobs. But as machines take over more and more knowledge jobs, this strategy is unlikely to be effective and certainly not for tens of millions of newly unemployed workers.
There is a possibility of a basic income to be provided to all citizens, to ensure basic subsistence. Nobel Prize winning economist Friedrich Hayek, held in high esteem by conservatives, proposed this. He believed that a guaranteed income was a legitimate government policy to provide insurance against adversity, where many individuals could no longer rely on traditional support systems as society transitioned. But, as the rich get richer, they are likely to wield greater political power and will resist attempts to redistribute wealth. And they do use friendly offshore jurisdictions to shield their wealth.
One factor that may be a mitigating factor is that workers are also consumers. If people do not have jobs, they cannot buy the greater output that will be created by robots. If corporations cannot sell their goods in sufficient quantities, then corporate profits may actually fall. This may lead to a compromise in how robots are used versus human labour, but it is not clear how this will resolve itself.
The rise of the robots will raise questions going forward. What is the role of wealth in society and how should it be shared? Given the taxes paid by all citizens to create the infrastructure and research that corporations benefit from, are they entitled to the wealth created by corporations? Will the middle classes allow the gains they have made in the last 70 years to be taken away or will a new medieval era come to pass where there is a small elite class and virtually everyone else is poor? What choice do people really have as these changes seem inevitable?
Oxford University Study on the Future of Employment
Top Twenty Affected (from 700 Job Types)
|Least Computerisable||Most Computerisable|
|Recreation Therapist||New Account Clerks|
|Supervisors of Mechanics, Installers, Repairers||Photographic Process Workers|
|Emergency Management Directors||Tax Preparers|
|Mental Health and Substance Abuse Social Workers||Cargo and Freight Agents|
|Occupational Therapists||Insurance Underwriters|
|Orthotists and Prosthetists||Mathematical Technicians|
|Healthcare Social Workers||Hand Sewers (Clothing)|
|Oral and Maxillofacial Surgeons||Title Examiners, Abstractors and Searchers|
|Supervisors of Fire Fighting and Preventions Workers||Telemarketers|
|Dieticians and Nutritionists||Library Technicians|
|Lodging Managers||Data Entry Keyers|
|Choreographers||Timing Device Assemblers|
|Sales Engineers||Insurance Claims and Policy Processing Clerks|
|Physicians and Surgeons||Brokerage Clerks|
|Instructional Coordinators||Order Clerks|
|Supervisors, Police and Detectives||Insurance Appraisers, Auto Damage|
|Dentists||Umpires and Referees|
|Elementary School Teachers||Tellers|