Two centuries ago David Ricardo had a change of heart. In 1819, testifying before Parliament, Ricardo — whose Principles of Political Economy has a better claim than Adam Smith’s The Wealth of Nations to be considered the founding document of the field we call economics — had dismissed concerns that the Industrial Revolution then in progress would hurt workers. But by 1821, when he published the third edition of his book, he admitted that his views “had undergone a considerable change.” He went on to declare that
the substitution of machinery for human labour, is often very injurious to the interests of the class of labourers
and that
the opinion entertained by the labouring class, that the employment of machinery is frequently detrimental to their interests, is not founded on prejudice and error, but is conformable to the correct principles of political economy.
In the language of modern economics, Ricardo had come to the conclusion that the Industrial Revolution was bringing about “capital-biased technological change”: change that, other things equal, reduced the demand for labor while increasing the demand for capital. And as he realized, strong capital-biased technology can drive wages down even as it raises GDP.
Let me note, by the way, that Ricardo was willing both to change his views in the face of new evidence and to admit that he had been wrong in the past. That kind of intellectual maturity and honesty is hard to find these days.
Was Ricardo right to conclude, after reconsidering, that the introduction of machinery was hurting workers? Economists are still debating that question, in part because economic data on early 19th-century Britain are limited. But Ricardo’s reversal shows that one of the biggest debates surrounding AI — will it hurt workers? — has a very long history.
AI itself, by contrast, has a very short history, and what we think we know about its economics is changing rapidly. My own views have evolved since the last time I wrote about this subject less than 4 months ago — hey, if Ricardo’s views could change, so can mine.
Today, then, I will offer another entry in my series of primers on issues related to AI, this one revisiting the question of how this powerful but confusing technology may affect incomes and the distribution of income going forward. Where Ricardo became more pessimistic about the potential impact of technology on the relationship between capital and labor, I am somewhat less pessimistic than I was a few months ago — or at least more skeptical about some of the extreme scenarios. I am, however, increasingly concerned about how AI will affect the reward or lack thereof for many traditionally valuable skills.
I’ll explain why beyond the paywall, where I will address the following:
· Will AI bring capital-biased technological change?
· How will AI affect the market for skill(s)?
