Firms are embracing artificial intelligence at a rapid pace, but they're not firing workers—at least not yet. A new analysis published August 5, 2026 by the Federal Reserve Bank of New York finds that while AI use jumped sharply across industries over the past year, companies report very few AI-driven layoffs. Instead, the technology is changing what skills employers need and how they recruit, with businesses planning to retrain workers rather than replace them. The findings offer a more measured view of AI's labor-market disruption than many workers fear.
The numbers show just how quickly AI has spread. Among service firms, 25 percent used AI in 2024, climbing to 40 percent in 2025 and projected to reach 46 percent within six months, according to the New York Fed's regional business surveys. Manufacturers lagged slightly, moving from 16 percent in 2024 to 26 percent in 2025, with 33 percent expected in the next half-year. Despite this surge, companies overwhelmingly plan to retrain employees for AI use rather than let them go, the data show. Recruiting patterns are shifting in two directions: some companies are cutting hiring because AI automates tasks, while others are expanding recruitment to find workers skilled in AI. Firms anticipate more reductions in hiring plans ahead, especially for college-educated employees.
The report's author, Kartik B. Athreya, research director at the New York Fed, notes that understanding AI's effects on labor markets is critical because the Federal Reserve has a mandate to conduct policy consistent with "maximum employment." The analysis found that back in October 2024, individuals who had been exposed to generative AI tools appeared to hold bleaker expectations for job availability and income inequality. Yet the business surveys tell a different story: "AI's labor-market impact has more to do with changing skill requirements than eliminating jobs—at least so far," the report states.
Why the disconnect between worker anxiety and employer behavior? Athreya explains that it's useful to think of each job as a "bundle of tasks" rather than a single function. While AI takes over some tasks, that frees up teams to focus on other areas of production and respond to shifting demands—a potentially positive aspect of the disruption. The real risk, he argues, is that today's economy relies on specialization: each person delivers only a narrow set of skills, whether as an economist, florist, architect, or welder. "If we had to do everything ourselves, as would be the case if we were marooned on a desert isle, we'd welcome AI with open arms," Athreya writes. The friction comes from our vulnerability to a sudden collapse in the value of the only skills we may have, making retraining and a stronger safety net worthwhile responses. He also points to AI's likely impact on the demand side of the economy, suggesting the technology will make many goods and services far cheaper, effectively making consumers richer.
Looking ahead, the report cautions that while AI's short-run effects may be more evolutionary than disruptive, certain occupations remain more vulnerable than others. The slow "absorption" or "diffusion" of new technology isn't unprecedented—the report cites the long lag between the invention of steam power and its replacement of traditional propulsion on American rivers. Athreya also raises a longer-term concern: as AI-enhanced algorithms supercharge social media platforms, spending habits could shift toward status-driven comparisons rather than need-based purchases, creating a wasteful zero-sum game. For now, though, the evidence suggests firms are preparing workers for new tasks rather than replacing them wholesale, even as anxiety about AI's ultimate impact lingers.