Recent advances in artificial intelligence (AI), particularly the generator AI, have led to an explosion of commercial and public interest in these technologies. High-tech companies invest billions of dollars
To expand AI capabilities, and millions of Americans use these technologies actively or passively every day. Both modern AI tools are powerful and can accomplish a wider range of tasks than their predecessors. Importantly, they are more accessible. Large-scale language models using electric tools like CHATGPT allow you to interact with these technologies using plain English without the need for advanced programming skills. AI optimists believe that it will ultimately lead to widespread productivity gains and reconstruct the future of work (although not all experts share this view).
What do you know about how recent AI developments have affected employment and how demand for AI-related skills in the labor market is affecting? Some of the best systematic evidence comes from online job posting data. The reason is that these technologies need to be developed and implemented by workers with a specific skill set, and used by workers who are familiar with them. Therefore, to some extent, job posting provides a signal about the demand for AI skills, which is explicit and to the extent that companies are looking for in their future employees. The fact that most vacant seats were posted online (according to some estimates) 
Coupled with constant termination in the labor market, approximately 60-70% of all vacancy (approximately 60-70%) will allow these signals to be acquired in real time.
This blog post draws on online job posting data from LightCast (formerly known as Burning Glass) to document how demand for AI skills has evolved. Lightcast continuously aggregates job ads from thousands of online job posting boards, thereby attracting jobs near online job postings. With each job offer, Lightcast records details such as employer, position, industry, location, experience or educational requirements. Importantly, Lightcast also extracts detailed skills listed in job descriptions. Define an “AI Job” as a job listing at least one “AI skill.” This is identified using the “AI and ML” skill category in Lightcast. This category includes 173 unique skills related to AI and machine learning (ML), spanning common terms such as “artificial intelligence,” specific concepts such as “neural networks,” programming packages such as “Pytorch,” and applications such as “ChatGpt.”
Online job postings are a selected subset of all vacant seats (some jobs are more likely to be posted online than others), but this data represents high-skilled jobs that are likely to demand AI skills very well. The long-term high frequency of these data allows us to document the changes in the demand for AI skills from 2010 to 31 August 2024, and the large nature of these data allows us to break down the demand for AI skills by detailed occupations and industries (data includes over 400 million unique job postings).
This blog post will document three main findings. First, demand for AI skills in the US labor market has been steadily rising since 2010, and this trend appears to have accelerated last year. Second, in recent years, the nature of the demand for AI skills has shifted from ML-related skills to AI-related skills. Finally, the demand for AI skills is spreading across a wider range of occupations, industries and local labor markets.
The demand for AI skills is increasing over time, and this trend appears to be accelerating
Figure 1 plots the share of online job openings for AI jobs. In line with previous research using the same data source (Alekseeva et al. 2021Acemoglu et al. 2022), we found that the demand for AI skills has steadily increased from 0.16% to 0.84% between 2010 and 2019. However, expanding the series to the present has revealed that last year appears to have been faster. Demand for AI skills fell slightly in 2023 (reflecting slower investments in pandemic post-miss in the technology sector), since then it has been rebounded in a big way. For the first eight months of 2024, the percentage of online AI Jobs vacancies was 1.62%. Although it is uncertain whether this trend will last, evidence suggests that demand for AI skills is accelerating.
There has been a recent shift from ML-related skills to AI-related skills
Examining the composition of demand for AI skills reveals some interesting patterns. Although AI and ML are difficult to separate, in recent years the demand for AI skills has shifted away from skills. ML is usually related to skills related to AI. Figure 2 shows this shift, extracting all AI skills into all AI jobs and plotting distribution across four categories. AI-related skills are skills that contain the term “AI” (such as “AI system”). ML-related skills are similarly defined, but also include “traditional” ML concepts such as “random forest algorithms” and “ensemble methods” (see Figure 2 below for more information). All other skills were conservatively assigned to the “Other/Organized Skills” category.
Between 2010 and 2016, ML-related skills accounted for around 40-50% of all AI skills listed in online recruitment, while the share of AI-related skills only amounted to around 10-20%. Since then, the balance has changed. Between 2016 and 2024, the share of AI-related skills rose to 45%, while the share of ML-related skills fell to 30%. The recent emergence of generator AI is also evident in Figure 2: Since 2022, the introduction of ChatGPT has reached about 5% share in the “generated AI” skill category, highlighting the AI field rapidly evolving and as a result, the type of AI skills in demand.
The demand for AI skills is spreading across a wider range of jobs, industries and local labor markets
Figure 3 shows the shares of occupations, industry and commuting zones (clusters of counties with strong commuting ties) with an AI online job share of at least 1%. Until 2015, the demand for AI skills was very concentrated. Since then, the demand for AI skills has spread across a larger occupational set, with almost a quarter of all occupations minimising the demand for AI skills by 2024. As these occupations are naturally distributed across more diverse industries and locations, the share of industries and commuting zones with minimal demand for AI skills has also increased in tandem, reaching 29% and 13% in 2024, respectively.
Which jobs increase the demand for AI skills? It is convenient to develop and implement AI technology, as well as distinguishing between computer and mathematical jobs that claim all the remaining jobs. Figure 4 shows the top 15 occupations for computer and mathematics occupations (Panel A) and all other occupations (Panel B) in terms of share of AI employment online recruitment in 2024. Corresponding stocks for 2010 and 2017 are shown for comparison. Panel A shows that until 2017, the demand for AI skills was concentrated in a small number of computer and mathematical professions, particularly computer scientists, data scientists and statisticians. However, the 2024 pattern suggests that AI skills are becoming increasingly common in most computer-related jobs. Perhaps one reason is that AI-powered coding tools can help enhance and accelerate programming tasks. All of these jobs are involved to varying degrees. Another likely reason is that the development and deployment of AI solutions requires a team of professional workers (to set up the required data storage and processing infrastructure, network functions, security protocols, and more).
It's not surprising that the demand for AI skills is increasing with high-tech jobs. what teeth However, many other jobs also have a rapid increase in demand for these skills. This is because modern AI technologies can (and are increasingly performing better) tasks, including text analysis, pattern detection, speech recognition, image processing, and data analysis. As a result, Panel B shows that the demand for AI skills is widespread across a wide range of occupations. The demand for such skills was previously low or virtually nonexistent. These include not only life sciences (including biological scientists, bioengineers, bioengineers, bioengineers, etc.) or business and finance (including economists, financial specialists, management analysts, and financial risk experts), but also writing (e.g., proofreaders, technical writers, authors, editors, etc.).
Overall, the pattern in Figure 4 highlights a number of applications of the latest AI technology. These findings are consistent with recent evidence. From the 2023-24 Census Bureau survey, where 27% of US companies used AI to perform tasks performed previously by workers and performed recent evidence. From a nationally representative household survey showing the widespread adoption of generative AI use (spanning a wide range of tasks and occupations, both at home and at work).
Conclusion
Demand for AI skills is growing in the broader segment of the US labor market. While it is still too early to determine whether AI technology will revolutionize the future of work, the evidence suggests that AI is beginning to have a meaningful impact on growing employment.
