This episode of The Intelligence from The Economist examines the hype around AI-driven job displacement by contrasting it with historical labor-market disruptions, particularly the Industrial Revolution and mid-20th-century technological changes. The discussion argues that historical evidence shows new technologies spread slowly, rarely causing mass long-term unemployment, and that current fears may be overstated due to Silicon Valley's historical ignorance and self-serving narratives.
Summarized by Podsumo
Historical analysis of the Industrial Revolution reveals that job destruction was minimal; employment in Britain nearly tripled in the century after 1760, and wage growth stagnated mainly due to slow productivity growth and rapid population growth, not technology.
Mid-20th-century technological disruption (computers, plastics, shipping containers) caused higher job churn than the Industrial Revolution, yet that era is now viewed as a golden age for workers, challenging the notion that rapid change necessarily harms labor markets.
Current AI fear-mongering by tech leaders may be driven by branding for IPOs and a genuine but flawed belief that everyone will adopt AI as they do, ignoring that productivity growth hasn't surpassed historical norms (2-2.5% per year) and the labor market remains strong.
"The average person believes they've got about a one-in-five chance of losing their job in the next five years. Another poll found a similar fraction thinks that AI is very or somewhat likely to replace them in that same time. Who can blame these pessimists? That's what the big shots of AI are telling them."
"There is a general rule looking back over a thousand years of history that per person GDP growth of the most technologically advanced country never exceeds about 2-2.5% per year. So if America were to exceed that, that would be one sign."
"I think a sort of corollary of Silicon Valley being so forward looking is that people there are actually quite ignorant of history and it's remarkable how little interest people have in previous technologies."