Will AI shorten the work week?

No. Sorry, stop crying. Still no.

I’ve seen a few postings by the oeconomica illiteratus arguing that the ability of artificial intelligence to increase worker productivity by 20 to 40% should translate into a 1 or 2-day shorter workweek (here and here). First, that’s their increase in productivity estimate, not ours. Second, France’s effort to legislate work hours aside, structural increases in productivity always result in a reallocation of resources, not a reduction in time. We have more than 1000 published ROI case studies that confirm that point.

“Greed is good.”

That iconic phrase from the movie “Wall Street” was based on a commencement speech by Ivan Boesky at UC Berkeley (wow, how times have changed there). I expect Oliver Stone intended it to be a condemnation, but it turned into a crystalizing economic truism. If you tell employees they can keep their salary and work a 3-day work week or work 5 days and get 66% more, well, it only takes a few to accept that deal to displace all of those 3-day workers. In very little time you reach a new economic equilibrium of fewer higher-paid workers delivering greater productivity per worker.

AI will replace (some) workers.

In 1914 Ford produced somewhere around 23 cars per employee per year. Today, Tesla can produce approximately 134 cars at its Freemont plant per employee per year (these numbers are close but I’m just using them to illustrate a point) If we applied all that increase in productivity from 1914 to today to only reduce the workweek, then the average UAW member would work fewer than 7 hours per week. Today’s farmers would work only a few minutes.

I won’t predict doom and gloom from the rise of artificial intelligence, but I will predict stratification. Employees that add value by using artificial intelligence to uncover and drive new ideas will thrive. Employees whose jobs are comfortably patterned may find a toughening future as AI starts to replace their more limited value.