The rise of artificial intelligence has brought fears of a potential collapse of the Science, Engineering, Mathematics and Technology (STEM) professional job market. In this month’s Race, Jobs and Economy update, we will look at the change in STEM employment for different racial groups and how these changes intersect with AI developments and other major economic events.
The number of STEM jobs has grown dramatically in the past decade. Many of these occupations have what researchers at Anthropic call “high AI exposure.” This means that their tasks and job functions have a high degree of overlap with the capabilities of artificial intelligence systems.
The impact of AI on the labor market and some STEM occupations has been widely discussed in the press lately. The Economist recently reported what they termed the “AI jobs boom.” They estimate that the number of certain STEM occupations like software development – jobs with the highest AI exposure rates – have increased at a faster rate relative to all other professional occupations. That cheery analysis ignores the decline in entry-level employment and how different racial groups have or have not benefitted from the AI jobs boom.
Before looking at STEM employment trends for racial groups, we should ground the analysis by examining employment changes in STEM industries and occupations overall. If we examine STEM industry jobs in the engineering, computer systems and scientific research and development (R&D) fields, job growth has stalled or gone negative.
The chart below plots the job totals for each sector indexed to their 2019 average:
The change from positive growth to a virtual slowdown can be seen in the scientific R&D industry. It has experienced a 30% increase compared to its 2019 average but this growth has tapered off since the beginning of 2024. Similarly, the computer systems industry has been slowly losing jobs since the spring of 2023. Coincidentally, the AI operating system ChatGPT 4 was introduced in March of that year.
Not all STEM occupations are represented in all STEM industries and not all workers in the STEM industry are STEM workers specifically. To better understand the fluctuating trends in STEM employment, we can look at the change in employment for STEM occupations. In the chart below, the employment totals for the three major STEM occupations indexed to their 2019 average are plotted:
What stands out is the proliferation of mathematicians and statisticians, a by-product of the expansion of the data science field. The other three STEM occupations have seen more modest growth, with computer and network occupations experiencing stagnation since 2023. The employment trends for STEM occupations as a whole are more or less in line with STEM industries overall – modest or stagnant growth.
With this background knowledge, we can examine the changes in STEM employment for different racial groups. Plotted below is the total number of employed STEM workers by race indexed to their 2019 average:
There are two noticeable trends. First, Asian and Hispanic STEM employment have reached their peak and risen dramatically since the start of 2025. Second, over this same time period Black and White STEM employment has declined slightly, though the latter is still 25% above its 2019 average. Both of these observations merit further analysis.
One explanation for the Asian and Hispanic STEM boom is immigration. Skills-based immigration pathways have been increasingly utilized by STEM workers. Foreign-born workers accounted for 19% of the STEM labor force in 2021. Of all foreign-born STEM workers, 18% were from Mexico, 14% from India and 7% from China.
Some of the increase could be due to the differences in STEM occupations held by Asian and Hispanic Americans compared to Black and White Americans. For example, 29% of statisticians, which according to the previous chart is a booming field, are Asian although they are only 6% of the population.
The decline in Black STEM workers since 2025 overlaps with larger DOGE cuts that gutted many government agencies, including the Center for Disease Control whose workforce was 27% Black in 2018. This development is not surprising given the DOGE cuts to federal employment, which was instrumental in building the Black middle class.
However, the decline in White STEM employment is perplexing and defies easy explanations. Though they are overrepresented in web development occupations, Asian Americans are overrepresented in the computer programming and software development fields. All three of these occupations fall within the broader computer and network group that has stagnated in recent years.
The retirement of the largely White baby boomer generation is another potential explanation. Moreover, some scientists chose to retire because of last year’s DOGE cuts affecting scientific R&D institutions. Additionally, some STEM graduates took jobs in the financial services sector where their occupations, such as becoming a quantitative analyst, could be defined as non-STEM. Nevertheless, the number of White STEM workers has declined in recent years but is still above their 2019 average.
STEM employment remains robust despite the potential of artificial intelligence systems replacing human workers. Growth has been sluggish in certain industries and occupations, particularly in computer-related fields. While the growth of Asian and Hispanic workers in STEM fields shows no signs of slowing, job growth for Black and White STEM workers appears to have stagnated in the last two years. While it is not apparent if these changes can be directly attributed to AI, they do show how simplistic narratives of an AI jobs boom in STEM deserve greater scrutiny with regards to who is and who is not benefitting.
Joseph Dean is the Junior Racial & Economic Research Specialist with NCRC’s Research team.
Photo credit: Edward Jenner via Pexels.
