Another one flowing from that assembly line of pieces mithering about what will happen once the robots steal all our jobs. And this is one area where we might hope that automation will indeed take over, and quickly too, because we should assume that machine learning will lead to the fully automated version being correct.
Automation does not lower the price of labour, automation raises the rewards to labour:
To sample what lies ahead, my colleague Ryan Avent in his book in automation’s impact, The Wealth of Humans,tried out a computer program that copied complex analytical journalism. Happily for my trade, the result was not a must-read. Yet even if we columnists are not yet to be displaced by some offspring of Sophia, with an expertly engineered prose style, it’s a safe bet that some parts of many jobs will be altered or eradicated by automation much faster than we reckoned on or find comfortable. A report for the World Economic Forum estimated that technology would create about 2m jobs by 2020 worldwide, but displace 7m.
For the educated and financially secure, this causes apprehension enough. Now imagine you are either without a job or in low-paid, patchy work, in economies facing the resulting efficiency gains but lowering the price of labour and wearing away workers’ bargaining power.
I think we'll all agree that the world is rather more automated than it was 240 years ago, when Adam Smith published WoT? And I think we'd all agree that labour is paid rather more today than then? Thus what is going to be our first pass attempt at portraying the relationship between automation and the rewards to labour? More means more, right?
In more detail we could take Paul Krugman's explanation. The average wage in an economy is determined by the average productivity of labour in that economy. Up goes the productivity and up goes the wage - automation increases labour productivity.
We can even dig right down into the details and consider why a barber in Nuneaton makes a higher income than a barber in Nairobi. Both are doing the same task using the same technology with the same level of productivity. But as William Baumol's cost disease argument points out, the higher productivity of the alternative uses of that labour pull up that barber's wage.
And it really is that alternative use of the labour which determines the general level of wages. And as automation makes the labour running the machines more valuable so does the real wage rate rise.
As it must do of course - because now we've the machines making myriad things which must be consumed, we humans are the only people who can do the consuming and thus we get to consume more. Our real wages have risen.