The history of technological innovations going back more than 200 years demonstrates that while certain industries endured major disruption, overall employment saw net gains. The AI revolution will be no different, according to JPMorgan.
Instead of creating mass unemployment, the breakthroughs slashed costs and gave rise to new functions that more than offset losses from obsolete jobs, he explained.
“We think AI could follow the same trajectory: Violent task churn, then broad productivity growth,” Manoukian wrote.
To be sure, workers performing old tasks have suffered significant losses. For example, hand-loom weavers saw their real wages cut in half between 1806 and 1820 after the introduction of the steam engine. Canal boatmen and cart drivers lost their jobs as steam-powered trains became the dominant form of transportation.
But as textile production and consumption soared while shipping costs fell, demand for labor in coal mining, rail maintenance and urban retail took off, he pointed out.
Electricity and computers spurred similarly broad transformations, many of which were unforeseen. And along the way, productivity boomed. Manoukian estimated that companies needed eight employees for every $1 million in revenue in the 1980s, but that shrank to six by the 2000s.
“This highlights another long-term trend: New innovations are contributing more quickly to overall productivity growth,” he added.
Given how quickly new technologies have been translating to productivity improvements, Manoukian thinks that even AI optimists are underestimating how fast the next boom will be.
After the steam engine was introduced, 61 years passed until productivity rose. That interval shrank to 32 years with electricity and 15 years with computers and the internet. For AI, JPMorgan estimates it will be less than seven years.
Anthropic CEO Dario Amodei has said AI could wipe out roughly 50% of all entry-level white-collar jobs within five years, causing overall unemployment to spike as high as 20%.
JPMorgan’s Manoukian sees some enduring advantages that humans have over AI, such as, common sense, causal inference, dexterity, emotional intelligence, high stakes accountability, adaptive learning and intrinsic motivation. AI could also help offset declines in the working-age population caused by aging demographics and stricter immigration policies.
Still, there are ways to mitigate the negative impacts from AI. For instance, the Federal Reserve can lower rates to stoke demand in areas like housing that are sensitive to borrowing costs. Policymakers can encourage apprenticeship programs. And employers can use AI to replace low-value jobs and retrain workers for new roles.
At the same time, businesses will likely reinvest some of the saving derived from AI into new growth areas, Manoukian said. So expect more hiring from companies that build software applications and data infrastructure as well as those that integrate AI tools into workflows and systems.
“The value proposition of AI is that it reduces costs and increases output for enterprises by making all workers more productive and specific tasks obsolete. Said differently, the total addressable market for AI is human labor,” he added.
“But the pattern we’ve shown, across generations of technological innovations, strongly suggests that AI will increase productivity and economic growth, and will create new channels for aggregate demand without lasting labor market damage. Disruption, not destruction.”
