Retraining Workers for “AI”: CHART OF THE DAY
No typing-pool annihilation this time, at least not so far, and probably never. AI isn’t killing occupations, but watch out for the elasticity of demand!:
Again, another one from Torsten Slok, who is on a roll these days:
With comment:
Torsten Slok: AI Is Retraining Workers, Not Replacing Them <https://www.apollo.com/wealth/insights-news/insights/daily-spark/ai-is-retraining-workers-not-replacing-them>: ‘A new survey from the New York Fed… shows that 34% of service firms and 22% of manufacturers using AI are retraining staff, while only 4% and 0% report layoffs. This is consistent with our core view that AI is putting downward pressure on wages in AI-exposed occupations without a significant negative impact on employment…. Here: Analysis of actual Claude usage data… suggest[s]… companies are capturing AI productivity gains through wage compression rather than workforce reduction…. [Analysis] by Sania Edlich and me using a difference-in-differences methodology with occupation and year fixed effects across 321 matched occupations from 2015 to today…
What are the jobs that are like the hand spinners, hand weavers, and hand knitters whose entire occupations vanished during the British Industrial Revolution of the 1800s that brought us into the SteamPower Age? What are the jobs that are like the typing pool stenographers and the switchboard operators and the back-office hand reconcilers whose entire occupations similarly vanished with the coming of the personal computer, = the electronic telephone switch, and the computer network?
Looking around, I see damned few of them.
Thus I think the way to bet is that tasks will shift and occupations will change, but whether the number of people employed in them depends overwhelmingly on the elasticity of demand for the kinds of things that they do. And so Jevons’s Paradox is key here, not as a totem and a fetish to wave around, but as an important piece of analysis.