Redefining AI jobs to steer policy

A new taxonomy finally counts who builds AI. Georgetown University’s Center for Security and Emerging Technology lays out a precise way to separate AI development work from related tasks, a shift policymakers have long debated. https://cset.georgetown.edu/article/defining-the-ai-workforce/

The framework defines AI development jobs as roles that directly contribute to the technical development of AI systems, not the broader spread of AI tools across organizations. In other words, it targets the hands-on builders, researchers, and engineers who literally create the models and infrastructure. https://cset.georgetown.edu/article/defining-the-ai-workforce/

The authors argue that official labor statistics do not neatly capture AI work, which has led to fuzzy counts and policy blind spots. They distinguish three distinct labor markets: people building AI systems, people adopting AI tools in other roles, and workers whose tasks are exposed to AI-enabled change. This nuance matters for both policy design and workforce planning. https://cset.georgetown.edu/article/defining-the-ai-workforce/

To make the data actionable, the taxonomy uses job postings data to measure demand for AI development roles, offering a way to gauge shortages, cost of talent, and training needs. The approach aims to provide a clearer signal for governments and firms trying to calibrate investments in education, visa policies, and industry incentives. https://cset.georgetown.edu/article/defining-the-ai-workforce/

Policy researchers cite concrete benefits from this precision: it helps determine where training pipelines are most needed, where recruitment may outpace supply, and how to allocate resources for AI capability across regions and sectors. The argument is that sharper definitions reduce misclassification and sharpen strategic planning, especially as AI moves from a speculative technology to a routine industrial tool. https://cset.georgetown.edu/article/defining-ai-workforce/

Yet the method also faces real-world constraints. Skill-based counts, the usual alternative, risk collapsing AI development work with AI-adjacent roles and adoption duties, muddying the policy signal. Critics warn that even a refined taxonomy can lag behind rapidly evolving AI tools and organizational structures, requiring ongoing updates to stay relevant. https://cset.georgetown.edu/article/defining-the-ai-workforce/

Practitioners watching the rollout will look for several next steps: aligning national statistics with the new taxonomy, standardizing job postings data across sectors, and building interoperability with international work-forces definitions. They will also monitor how the framework affects education programs, visa and immigration considerations for AI specialists, and corporate talent strategies. All of these hinge on keeping the definitions tight enough to be actionable while flexible enough to capture fast-moving technical work. https://cset.georgetown.edu/article/defining-the-ai-workforce/