Skill Gap Analyses
Task-level measurement frameworks that identify which occupations and communities face the greatest AI exposure, directing retraining resources where they are most needed.
What it is:
Skill-gap analyses for AI apply task-based frameworks that break occupations down into their component tasks, then assess which tasks are susceptible to automation or augmentation by AI systems. Rather than predicting wholesale job elimination, this approach recognizes that AI affects work at the task level; some tasks within an occupation may be highly exposed while others remain resistant to automation.
Skill-gap analyses offer policymakers a more precise alternative to broad-based workforce interventions. By identifying which tasks, occupations, and communities face the greatest exposure to AI, these frameworks can help direct retraining funding where it is most needed, enable community colleges to anticipate demand for emerging credentials, and allow employers to distinguish between roles that are candidates for augmentation and those likely to shrink. Importantly, high task exposure does not automatically translate into job displacement; the same analysis that identifies automatable tasks may also reveal opportunities for workers to shift toward higher-value activities within the same occupation.
The challenge:
The primary challenge is that task-based frameworks depend heavily on the quality and currency of the underlying occupational data. O*NET and similar occupational taxonomies are updated infrequently relative to the pace of AI capability development, meaning exposure estimates can quickly become outdated. Moreover, whether theoretical task exposure translates into actual labor-market outcomes depends on several factors, including how jobs are structured, regulatory constraints, the availability of retraining, and whether AI functions as a substitute for or complement to existing skills.
Recommended Reading:
Real-world precedents:
The U.S. Department of Labor's Occupational Information Network (O*NET) provides a foundational taxonomy, cataloging approximately 20,000 task statements across 900+ occupations. Researchers classify these tasks by their exposure to AI capabilities, producing occupational exposure scores that can inform workforce development priorities.
BLS's Employment Projections program incorporates AI-related impacts for occupations where high exposure to automation is deemed likely, and released a suite of skills data tables providing information about skill importance across occupations.
At the state level, skill-gap measurement is becoming embedded in workforce planning.
Tennessee's AI Workforce Action Plan draws directly on MIT's Iceberg Index to identify county-level exposure patterns and target training investments.
North Carolina State Senator DeAndrea Salvador, who worked closely with MIT on the project, notes the tool enables policymakers to examine county-specific data to identify skills likely to be automated or augmented.
Utah's Office of AI Policy is preparing a similar state-level analysis.