Today’s cautious hiring environment likely reflects more fundamental macro drivers: policy uncertainty, constrained labor supply and slowing economic growth.
The recent slowdown in U.S. hiring has renewed questions about AI and its impact on the workforce. Recent employment data shows job growth averaging just 35,000 from May through July, a significant drop from the 127,000 pace earlier in the year. Meanwhile, this earnings season has seen companies showcase a variety of AI strategies, from agentic tools that can write code or create PowerPoint presentations, to individualized pricing strategies for flights. Some CEOs have even directly signaled that AI may reduce their workforce in the coming years.
Is AI behind the recent hiring slowdown, or are we mistaking correlation for causation?
Still early in the AI adoption cycle
AI adoption remains in its infancy. Census data show that less than 10% of firms use AI to produce goods and services today, and while that data does not reflect intensity of AI usage, we know that true automation often requires integrated, domain-specific AI agents, governed by compliance systems and orchestrated by human teams. A recent McKinsey survey found that only 1% of businesses have mature AI deployments driving business outcomes.1
Even among companies that have access to tools like ChatGPT (and less than half subscribe to paid versions), general-purpose LLMs go nowhere near achieving wholesale workforce automation. Research shows that in the software engineering field, even the most advanced models achieve an 80% success rate in autonomously completing tasks that would require up to 20 minutes of human effort2—surely helpful, but far from replacing a full workday.
AI is reshaping roles, not eliminating them
In 2025, many companies have focused on retraining and upskilling in response to AI integration. IBM reported that AI-powered learning platforms reduced employee training time by up to 25% and improved engagement. Firms like PwC and MasterCard have used internal AI academies and expanded demand for AI-literate talent across business functions. In software engineering, tech CEOs emphasize the shift from code writing to managing AI-generated code, turning developers into “creative directors of code”. Despite significant AI exposure, the BLS projects an 18% increase in employment for software developers by 2033, much faster than the average for all occupations (4%), as AI drives demand for database administrators and architects to support AI solutions.
If AI isn’t the full story, then what is? Today’s cautious hiring environment likely reflects more fundamental macro drivers: policy uncertainty, constrained labor supply and slowing economic growth. As the policy picture crystallizes, hiring activity could improve. Indeed, as tariff turmoil subsided in June, the seasonally adjusted unemployment rate for recent college graduates improved to 4.8% from May’s 5.3%.3
A long-term transition is still ahead
Looking ahead, AI will become a more material driver of labor market change. The World Economic Forum expects that roughly 60% of jobs in advanced economies may be affected by AI, and about 39% of current workers’ skills will become outdated by 2030. AI will drive both job creation and displacement, with the net impact depending on how quickly economies and individuals adapt to new skill demands.
In the meantime, today’s hiring slowdown looks more like a reflection of economic caution than technological disruption.