AI Workforce Gets a Precise Definition

AI development jobs finally get a precise definition.

Georgetown University’s Center for Security and Emerging Technology (CSET) has rolled out a new approach to defining the AI workforce, aiming to cut through a murky landscape where terms like AI do not mean the same thing to everyone. The core of the proposal is simple in concept but potentially powerful in policy: AI development jobs are those roles that directly contribute to the technical development of AI systems, a distinction CSET says has been missing in official statistics and workforce surveys.

In practice, the taxonomy separates builders from AI-adjacent work. It also introduces a method for measuring demand for development roles by analyzing job postings data. The goal, according to the researchers, is to give policymakers a clearer picture of where talent gaps lie, how fast they are changing, and what kinds of training or immigration policies might be necessary to keep pace with AI progress. The problem the approach seeks to solve is familiar to many government strategists: official labor statistics do not neatly capture AI work, and the result is blurred signals about demand, missing shortages, and misaligned education pipelines.

The authors note that the term AI has become a catch-all used to describe a broad and evolving set of activities. Without a precise taxonomy, it is easy to overcount or undercount the people who actually build AI systems. By focusing on roles that directly contribute to technical development, the framework aims to provide a more stable basis for planning and investment. Policy documents that aim to grow domestic AI capabilities or to tailor education and training programs can use this clearer lens to decide where grants, apprenticeships, and curriculum updates belong.

Industry observers welcomed the effort as a practical tool for planning and risk assessment. A clearer map of AI development talent helps companies forecast hiring needs, identify bottlenecks, and justify investments in in-house upskilling or targeted recruitment from abroad. It also matters for visa and immigration policy, where competition for specialized technical talent has become a global hotspot. If governments can reliably identify which roles constitute AI development work, they can design more focused pathways for skilled workers and align funding to those priorities.

Educators and researchers see room for alignment between curricula and actual development tasks. The taxonomy could steer AI-related degree programs, certificate tracks, and hands-on training toward the skills most tightly linked to building AI systems. But there are caveats. Job postings data, the proposed measurement tool, can be incomplete or biased toward larger firms that publish openings widely. Some critical roles may be outsourced, classified under broader teams, or filled through internal promotions rather than advertised externally. The fast pace of AI change means the taxonomy will need periodic updates to stay current with new techniques, tools, and workflows.

Looking ahead, policymakers will want to watch how this taxonomy interacts with existing data ecosystems and international benchmarks. If adopted at scale, it could recalibrate how governments measure AI capacity, shape investment programs, and monitor progress over time. The real test will be whether the taxonomy can be integrated into national statistics, cross-country comparability, and ongoing labor-market surveillance in a way that does not lag behind the technology it seeks to measure.

As a practical takeaway, executives should consider how their hiring and training plans align with a clearly defined AI development footprint. Firms that already track complex AI roles across teams may find the taxonomy validates their approach, while others may discover hidden gaps between perceived and actual development work. In either case, the move signals a trend toward more disciplined talent planning in an era when the speed of AI progress increasingly hinges on having the right people in the right roles at the right time.

Sources

Defining the AI Workforce