What if AI is creating more jobs than it kills?

The layoff headline writes itself. The hiring headline takes more work — which is why it gets less airtime.

Look past the anxiety cycle and a different labour market is forming. Roles that barely had LinkedIn taxonomies five years ago now have headcount, budgets, and career ladders. Forward Deployed Engineers translate models into messy enterprise reality. Evaluation engineers score whether the system is lying. AI operations teams keep the lights on when the demo ends.

The constructive case

Every general-purpose technology creates complementary work. Electricity needed electricians. Spreadsheets needed analysts who could think in grids. Generative systems need people who can specify intent, audit outputs, and redesign processes around imperfect automation.

That is not cope. It is how technological employment usually works — with a lag, with uneven distribution, and with real pain for the displaced. Honesty requires both sentences.

Where the upside is loudest

• Roles that sit between model and customer: FDE, solutions, applied research that ships.
• Trust labour: red-teaming, policy, safety, compliance that actually reads the logs.
• Craft that agents amplify rather than replace: editors who direct, designers who constrain, teachers who redesign assessment.

The useful question is not “will AI take jobs?” It is “which new jobs are we staffing with the same seriousness as the ones we fear losing?” The organisations answering that second question are hiring. Quietly. On purpose.