In the debate over AI and employment, Boston University law professor James Bessen set the tone with a 2015 paper for the International Monetary Fund (IMF). Bessen argued that technologies do not destroy jobs; instead, they transform or reshape them. His central example was the impact of ATMs on bank tellers. Contrary to early predictions, the spread of ATMs coincided with an increase in the number of tellers. He offered two main explanations. First, ATMs lowered the cost of running branches, enabling banks to open more locations. Second, tellers shifted into broader, more complex roles as relationship managers, handling tasks that ATMs could not, from serving existing customers to attracting new ones. Historically, Bessen contends, this pattern has been common with most major technologies. This article appears in Branding Strategy Insider’s FREE newsletter. Join leading marketers worldwide and subscribe here to receive practical insights straight to your inbox. Bessen’s paper is now cited in nearly every serious conversation about AI and work. A quick Google search for “ATMs and jobs” turns up analyses from the American Enterprise Institute, Brookings, the Hoover Institution, The Economist and the American Economic Association, all referencing Bessen’s findings approvingly. Yet it may be that Bessen’s analytical window was too narrow. That’s the argument of VC David Oks in a recent Substack post. While Bessen examined the period from 1970 to 2010, Oks extends the data through 2025. And over the last 15 years, the number of tellers has…