The Stanford Digital Economy Lab’s Canaries Dashboard has been updated to reflect labor market trends following the introduction of generative artificial intelligence, particularly ChatGPT, in November 2022. Research by Brynjolfsson, Chandar, and Chen (2025) indicates that workers are potentially at risk of displacement or growth based on their occupation and experience. This creates specific segments in the job market that can serve as indicators of broader employment trends.

Read More

The rapid integration of generative AI has ignited discussions about its effects on the workforce, ranging from increased productivity to concerns over job loss. Historical patterns reveal that technology impacts various occupations and industries differently, sometimes replacing jobs, augmenting others, and transforming still more. Consequently, certain occupations may act as early warning signs of AI’s broader impacts.

In examining employment trends since the launch of ChatGPT, data show that while all groups have seen employment growth, the growth rate is slowest for the most AI-exposed roles. For instance, among early-career workers (ages 22-25), the two most exposed occupation groups exhibit noticeable employment declines, in contrast to other groups that are growing.

Additional findings indicate that early-career software developers and customer service representatives, both in high-exposure categories, are experiencing significant job losses. Conversely, home health aides, which belong to less exposed groups, are observing job growth among younger workers.

Significantly, employment trends differ by age and gender as well. For early-career women, employment has contracted at a faster rate compared to their male counterparts within the highest exposure quintiles. Data from November 2022 reveal that the most exposed occupations constituted 38.3% of employment in the Canaries sample, which included 25,000 firms and 4.6 million matched workers.

The Dashboard utilized low-latency data provided by ADP, offering timely insights into employment across over 730 unique occupations. It serves as a tool to measure potential shifts in the labor market, informing efforts to address the challenges posed by AI.