September 30, 2026, (Inside AI) — Artificial intelligence can already perform a wide range of work tasks, yet the predicted wave of mass layoffs has not materialized. Economists and management experts say the delay stems from clumsy corporate implementation, the enduring value of human tacit knowledge, and a profit motive that has not yet aligned with full automation.
According to a National Bureau of Economic Research working paper published this spring, the labor market impact of AI remains highly uncertain. The study, co-authored by Harvard Kennedy School economist Doug Elmendorf, former director of the Congressional Budget Office, along with Karen Dynan and Louise Sheiner of the Brookings Institution, outlines four scenarios ranging from moderate GDP growth with little job loss to rapid growth with persistently high unemployment.
"What we've seen so far in the labor market from artificial intelligence has very little predictive power for what we're going to see in the labor market because of AI in five years, or 10 years, or 15 years," Elmendorf said.
The co-authors declined to assign probabilities to their scenarios, but Elmendorf expressed measured concern. "I think it's very likely that there will be significant job losses from artificial intelligence in the coming couple of decades, but that most -- not all, but most -- of those people will be able to find new work," he said.
Read: HBR Research Reveals AI Intensifies Work and Erodes Judgment
Their paper compares AI's potential disruption to the China shock of the early 2000s, which cost an estimated 1.5 to 2 million U.S. jobs between 2000 and 2007. The authors' AI scenario projects 3 million people, roughly 2 percent of the labor force, out of work at any given time, translating to many millions over decades.
Why Firms Are Stalling on Automation
Joseph Fuller, a professor at Harvard Business School, points to corporate ineptitude as a key brake on AI-driven layoffs. In collaboration with Accenture Research, Fuller developed a model analyzing work tasks across all job categories. It found that 41 percent of all work tasks can be automated or augmented by AI today. Yet only about one-third of firms' AI experiments succeed.
"Companies are not training people formally, so employees are basically using it like glorified Google, and they're not getting much productivity out of it," Fuller said. "Maybe the boss says, 'Well, this isn't as good as I was hoping, I guess we shouldn't worry too much about this.'"
Fuller warns this reprieve is temporary. He expects adoption to accelerate as cheaper, more specialized AI agents enter the market.
Raffaella Sadun, a professor at Harvard Business School and principal investigator at the HBS Digital Reskilling Lab, cautions against viewing AI as a replacement for human capital. "For the first time we have a technology that learns, sometimes autonomously, that allows us to have access to expertise that has previously been dispersed," she said. "I think that's a tremendous trap, a terrible trap."
Sadun argues that automating workers away trades short-term savings for long-term losses. Tacit knowledge, the firm-specific context employees hold, makes them more valuable than job descriptions suggest. At the Digital Reskilling Lab, she and co-principal investigator Jorge Tamayo retrain workers for AI-augmented roles and test the results.
Read: Employees Override AI Systems Even When They Work, Harvard Professor Says
"At the lab, we see investments in human capital and training as strategic investments, meaning these are things you should do not because you want to be nice or corporate-social-responsibility-compliant, but because knowledge plays an important role in your company, and investment in human capital is part of how you create knowledge," Sadun said.
Fuller works with executives to rethink entry-level roles. At one large tech company, he helped study senior engineers' performance reviews and found many lacked sales and product management experience. "So rather than stopping hiring junior software engineers, we changed the job description so those hires are still doing some software engineering, but they're also getting regular structured introductions to sales and product management," he said. "We can pre-position what we want 10 years from now by getting them the experience their current bosses didn't get."
Elmendorf remains skeptical that voluntary corporate restraint will hold. "Individual companies are going to be focused on what's good for their company, and yelling at them to do something different can matter a little bit but it won't fundamentally change their desire to produce goods and services at the lowest possible cost," he said. "And for a number of companies, that will mean the use of AI instead of workers for some things."
He argues that government must address collective-action problems the market cannot solve. The NBER paper suggests policies including wage insurance, subsidized private employment, shorter work weeks, and expanded retraining. "Policymakers should be prepared for much bigger economic changes than we've seen so far from AI," Elmendorf said. "Not because we're sure those changes will happen, but for the same reason you buy insurance against possible problems in your personal life: It might happen, and you'd rather not be surprised when it does."
For now, the unemployment rate hovers around 4.1 percent, buoyed by healthcare hiring. Accenture reports that 52 percent of top executives expect AI to increase entry-level hiring. The gap between dire forecasts and current reality may simply reflect the messy, uneven process of technological adoption, a pattern seen in previous industrial transitions.
