Every major technological wave has destroyed existing job categories and created new ones. The printing press eliminated hand-copyists and created typesetters, editors, and journalists. The automobile eliminated blacksmiths and created mechanics, traffic engineers, and urban planners. AI is following this pattern — eliminating routine task-based roles while generating demand for new competencies and entirely new job categories.
New roles emerging from AI adoption include: AI trainers and evaluators (who maintain model quality), AI ethicists and governance specialists, prompt engineers, AI integration consultants, machine learning engineers, AI product managers, and domain experts who serve as human-in-the-loop quality controllers for AI systems in their field.
Beyond AI-specific roles, productivity gains from AI adoption historically expand demand in adjacent fields. When automation makes a service cheaper, demand for that service grows — potentially creating more total jobs even after displacement. Legal AI tools are making legal services cheaper, which is expanding access to legal markets (creating new demand for legal professionals) while reducing demand for routine legal tasks.
The challenge for displaced workers is that new job creation happens in different roles, industries, and locations than displacement. A displaced bookkeeper does not automatically become a machine learning engineer. The transition requires reskilling, geographic mobility, or leveraging transferable skills toward adjacent roles — which is why policy attention to transition support is as important as the aggregate job creation numbers.