Ever since textile workers in 19th century Britain attacked the machines that eliminated their jobs technological disruption is met by society with suspicion and ambivalence. As the era of machine intelligence and human-like robots dawns economists scramble to determine the benefits and costs of the “second machine age”. Steam delivered the automation of manual labour; silicon, electronics and artificial intelligence are automating the intellect. Robots will no longer be confined to the assembly line, but will begin to invade the finance and legal departments, and soon design and engineering, sales and marketing, and all the way up to the C-suite.
Pessimists see almost half of the current jobs eliminated in the next ten years, resulting in millions of unemployable humans, social unrest, and the disintegration of democratic order. Optimists point to the flatlined productivity in developed economies over the past decade, and argue that only smart machines can save us from economic stagnation: exploiting the bounty of big data, automating tedious primary analyses, and freeing as well as augmenting human creativity, would usher a new industrial revolution, they argue. Curiously, both camps seem to agree that whatever happens – bull or bear – income inequality will continue to increase. “Average is over”, economist Tyler Cowen suggests in his book, arguing that only those trained to work with intelligent machines may thrive, while the rest will survive in near-subsistence, dependent on social welfare or its 21st century equivalent: “universal income”. So is this what the future has in store for the human race? A dystopia where most of us, replaced by robots, will squander our days without aim, purpose, or hope?
Economics has a long tradition of consistently failing to predict anything of substance, so we must approach economic crystal-gazers with extreme caution. Nevertheless, a recent study by McKinsey took a promising new approach in its analysis. Instead of considering occupations as the units of elimination by automation they went deeper and identified activities within occupations that could be performed better by intelligent machines. The study’s preliminary results suggest that only 5% of current occupations may be fully automated with present-day technology; however, the 60% of occupations that will remain in existence will have around 30% of their activities automated. If this productivity gain is passed solely to a company’s assets then – in the US alone – around two trillion dollars of annual wages will be lost. If the economy were a zero sum game, capital investment in cognitive automation would result in the owners of capital becoming richer due to the astronomical savings in labour cost, while the workers become poorer as their wages dwindle.
But the economy is not a zero sum game. For instance, the potential for transforming business processes due to cognitive automation can have benefits that are multiples of these wage losses. Furthermore, automating human cognitive tasks would act as a catalyst for an unpredictable torrent of microeconomic, as well as macroeconomic, transformations – which is why predicting the future based on current assumptions and logical extrapolations, is simply impossible. There are simply too many parameters to include in any predictive mathematical model. Will globalisation continue to open up markets? Will rapid innovation make the sharing economy the norm? How will human networks evolve, and how will they undermine the idea – or raison d’etre – of a public company? How will investment capital behave when governments refuse bail-outs, and boom and bust cycles increase in frequency? And, in an era of increasing international tax scrutiny, what legislation will be enacted to dampen income inequality?
Income inequality becomes a political issue when the spread of wealth widens while the quality of life for the majority worsens. If rapid innovation in science and technology benefits the majority then hoi polloi are likely to tolerate the robots that replace them, as well the rich (who invest in the robots) becoming richer. But what if the benefits serve only the few? With the many not having any meaningful disposable income to participate in economic activity, who will consume the outcomes of increased productivity?
This is the paradox of automation, which has been very eloquently articulated by an apocryphal conversation that purportedly took place in the 1950s between Henry Ford II and the leader of the automobile workers union Walter Reuther. The two rival men, it is said, were touring a newly-built and highly-automated factory, with shiny robots lining the assembly hall; when Henry Ford mused: “Walter, how are you going to get those robots to pay your union dues?”
Whereupon Walter Reuther replied: “Henry, how are you going to get them to buy your cars?”