New evidence from Wolfe Research, presented in the firm's recent "Roth Report," paints a nuanced picture of how artificial intelligence is reshaping the U.S. labor market. While early alarms have focused on mass displacement, the most recent data indicate that job creation has, up to this point, exceeded job losses.
Specifically, the report documents approximately 700,000 jobs lost to date in sectors exposed to AI, contrasted with an estimated 1 million positions created over the last two years. That gap produces a headline net gain, but it also obscures an important change in the make-up of employment.
Composition versus count
Analysts highlight a growing mismatch between the kinds of work being eliminated and the kinds of roles being created. Reductions are concentrated in routine positions within technology and finance, while hiring is shifting toward highly specialized roles focused on building, managing and auditing AI systems. The new jobs include titles that were rare or non-existent five years ago - for example, AI Ethicists, Algorithm Auditors and Prompt Engineers - and are likely to be a more technical and narrowly skilled cohort.
Wolfe Research notes that as many as 5 million jobs could remain at risk over the next decade. The firm frames the issue not simply as a question of net job counts but as a problem of whether the workforce can retrain at a sufficient pace to fill the technical roles that employers demand.
Sectoral patterns and recent labor data
The report finds early signs of this mismatch in recent labor-market performance. Cyclical industries such as construction and manufacturing have shown resilience, while hiring in traditional technology hubs has weakened. That divergence suggests the substitution effect of automation is uneven across the economy.
Investor and corporate implications
From an investor perspective, the transition signals a potential long-term productivity boost driven by automation. At the same time, Wolfe Research warns of short-term frictions: competition for a limited pool of specialized talent could produce "non-linear" wage volatility as firms bid for expertise in building and overseeing AI systems.
In practical terms, the report argues that the speed and effectiveness of retraining efforts - both within firms and through broader workforce initiatives - will be a central determinant of how smoothly the market adjusts. For corporate earnings and broader economic stability through 2026 and beyond, the capacity to bridge the skills gap may become a defining factor.
Takeaway
The early phase of AI adoption has produced both displacement and creation, with job creation currently outpacing losses. But the underlying transition is marked by a shift in job composition that poses material risks if workers cannot move into newly emerging, more technical roles. How quickly that adjustment occurs will shape labor-market outcomes and could influence wage dynamics and corporate performance as the decade progresses.