AI-Impacted Occupations Continue to Decline: Analyzing the 2024–2025 Employment Trends
In a striking contrast to the overall growth of the U.S. labor market, employment in occupations widely considered to be at high risk from artificial intelligence has continued its downward trend. According to fresh data from the Bureau of Labor Statistics (BLS), total employment across 18 job categories identified as AI-exposed dropped by 0.2% between May 2024 and May 2025. Over the same period, the broader American workforce expanded by 0.8%, signaling a widening gap between AI-vulnerable roles and the rest of the economy.
Key Findings from the Bureau of Labor Statistics
The latest figures, analyzed by Bloomberg’s Matthew Boesler, underscore a persistent pattern: for the second consecutive year, several occupations that are expected to be heavily influenced by artificial intelligence experienced notable job losses. The data set, released by the BLS, covers the 12-month period ending May 2025.
Overall Employment Trends
The national labor market demonstrated steady resilience, adding 0.8% more jobs from May 2024 to May 2025. Growth was driven by sectors such as healthcare, hospitality, and professional services, which continued to rebound from earlier fluctuations. This expansion brought total nonfarm payroll employment to record or near‑record levels, providing a broad backdrop of economic strength.
Decline in AI-Exposed Occupations
In contrast, the 18 occupations that researchers and analysts have flagged as particularly susceptible to automation and AI adoption saw a net reduction of 0.2% in their workforce. These roles—ranging from data entry clerks and telemarketers to customer service representatives and administrative support positions—have faced mounting pressure as companies integrate machine learning, natural language processing, and robotic process automation into daily operations.
The BLS does not officially label occupations as “AI-exposed,” but the list used in this analysis is based on prior academic and industry studies that rank jobs by their susceptibility to AI-driven displacement. The consistent contraction over two consecutive years suggests that the influence of artificial intelligence on these specific job categories is not a one-time shock, but an ongoing structural shift.
Which Occupations Were Hit Hardest?
While the BLS report does not break out every individual occupation, employment data for several of the 18 AI-exposed categories showed significant declines. Among the hardest hit were positions involving routine clerical and administrative tasks where AI-powered software can now perform much of the work faster and cheaper than human employees.
- Data entry keyers – Employment continued to drop as automated data capture and optical character recognition technologies improved.
- Telemarketers – With the rise of AI-driven voice bots, this role saw one of the steepest percentage losses.
- Customer service representatives – Many basic inquiries are now handled by chatbots and virtual assistants, reducing demand for human agents.
- Bank tellers and related clerks – Self-service kiosks and mobile banking apps further eroded these positions.
- Paralegals and legal assistants – AI document review and contract analysis tools have begun to replace some of their traditional duties.
Not all AI-exposed occupations declined; a few, such as computer support specialists, saw modest gains as companies needed humans to manage and oversee the same AI systems that displaced other roles. However, the overall trend remained negative.
Implications for Workers and Policymakers
The divergence between overall labor market growth and the contraction in AI-exposed jobs carries important implications. For workers, it highlights the urgency of upskilling and transitioning into roles that are less routine and more creative or interpersonal—areas where humans still hold a comparative advantage over machines.
Reskilling and Education
Demand is rising for skills in AI system management, data analysis, and roles that require emotional intelligence or complex problem-solving. Policymakers and educational institutions are under pressure to design training programs that prepare displaced workers for these new opportunities. Without such initiatives, the gap between “winning” and “losing” occupations may widen.
Broader Economic Effects
While aggregate employment has grown, the ongoing displacement in AI-exposed sectors could drag on wage growth and exacerbate income inequality. Industries that rely heavily on these roles may face higher turnover and costs as they automate. Meanwhile, the productivity gains from AI could eventually boost overall output, but those benefits may not be evenly distributed.
Comparison with Previous Years
The 0.2% decline in 2025 follows a similar drop in the previous 12-month period (May 2023 to May 2024), indicating that the effect is not a fluke but a developing trend. The consistency suggests that businesses are not just experimenting with AI but are making permanent changes to their staffing models. As AI technology continues to improve, the list of vulnerable occupations may expand, and the rate of job loss could accelerate.
However, economists caution against overinterpretation. The 0.2% decline is relatively small in absolute numbers—roughly 10,000 to 20,000 jobs—and the broader labor market added millions over the same period. Yet the symbolic weight is clear: AI’s impact on employment is no longer hypothetical; it is visible in the official statistics.
Conclusion
The BLS data for the year ending May 2025 provides compelling evidence that AI-exposed occupations are losing ground even as the overall U.S. job market flourishes. With a second consecutive year of contraction, these roles—ranging from clerical to customer service—appear to be undergoing a structural transformation. For workers in affected fields, the message is clear: adaptation is not optional. For employers and policymakers, the data serves as a reminder to invest in training and social safety nets that can help individuals navigate this new landscape.
Source: Bureau of Labor Statistics (May 2024 – May 2025 data, as reported by Bloomberg)