US labour market defies AI doom predictions as data reveals stable employment

Contrary to widespread predictions of an imminent jobs apocalypse, economic analysis of data from the US Bureau of Labor Statistics (BLS) indicates that artificial intelligence has not yet caused large-scale disruption to the American labour market. Unemployment rates for occupations most exposed to AI are currently lower than those for less exposed roles, contradicting fears that the technology is rapidly decimating white-collar work.

While the current statistics do not preclude future upheaval, they suggest that the pace of disruption is slower than anticipated. Erika McEntarfer, a labour economist and former head of the BLS, notes that AI’s impact on current conditions is likely small. She points to US Census data showing that only one in five companies are using AI in any business function, arguing that it takes time for innovations to transform industries and occupations.

Specific impacts are visible in entry-level sectors. A study by the Stanford Digital Economy Lab using ADP payroll data found a 16 per cent decline in entry-level jobs in AI-exposed occupations after 2024, particularly among workers aged 22 to 25 in software development. However, overall employment in coding jobs continues to grow, albeit at a slower pace. Economists at the Federal Reserve Board found that annual employment growth for coders has slowed by about 3 per cent since the introduction of ChatGPT, but the total number of coding positions remains on an upward trajectory.

Experts from Stanford University and the Federal Reserve note that while AI is transforming specific tasks, the technology has not yet replaced the tacit knowledge held by experienced workers. Wages in sectors highly exposed to AI have risen relatively fast since the launch of generative AI tools, suggesting employers are still willing to pay for the experience and judgment that older workers possess. This trend indicates that the traditional earn-while-you-learn model may be breaking down for some roles, rather than jobs disappearing entirely.

Interest is rising in AI-adjacent fields such as data science and cybersecurity, with artificial intelligence itself becoming a fast-growing college major. Researchers emphasise the need for better data to understand the pace of transition and the potential need for workforce reskilling. David Deming of Harvard University, who has been surveying workers since 2024, found that generative AI is used by a little over 40 per cent of workers, highlighting the importance of granular data to guide policy and prepare for future changes.